Thiol Metabolism and Redox Regulation of Cellular Functions (NATO: Life and Behavioural Sciences, 347) (Nato: Life and Behavioural Sciences, 347)

THIOL METABOLISM AND REDOX REGULATION OF CELLULAR FUNCTIONS

NATO Science Series A series presenting the results of scientific meetings supported under the NATO Science Programme. The series is published by IOS Press and Kluwer Academic Publishers in conjunction with the NATO Scientific Affairs Division. Sub-Series I. II. III. IV. V.

Life and Behavioural Sciences Mathematics, Physics and Chemistry Computer and Systems Sciences Earth and Environmental Sciences Science and Technology Policy

IOS Press Kluwer Academic Publishers IOS Press Kluwer Academic Publishers IOS Press

The NATO Science Series continues the series of books published formerly as the NATO ASI Series. The NATO Science Programme offers support for collaboration in civil science between scientists of countries of the Euro-Atlantic Partnership Council. The types of scientific meeting generally supported are "Advanced Study Institutes" and "Advanced Research Workshops", although other types of meeting are supported from time to time. The NATO Science Series collects together the results of these meetings. The meetings are co-organized by scientists from NATO countries and scientists from NATO's Partner countries - countries of the CIS and Central and Eastern Europe. Advanced Study Institutes are high-level tutorial courses offering in-depth study of latest advances in a field. Advanced Research Workshops are expert meetings aimed at critical assessment of a field, and identification of directions for future action. As a consequence of the restructuring of the NATO Science Programme in 1999, the NATO Science Series has been re-organized and there are currently five sub-series as noted above. Please consult the following web sites for information on previous volumes published in the series, as well as details of earlier sub-series: http://www.nato.int/science http://www.wkap.nl http://www.iospress.nl http://www.wtv-books.de/nato_pco.htm

Series I: Life and Behavioural Sciences - Vol. 347

ISSN: 1566-7693

Thiol Metabolism and Redox Regulation of Cellular Functions Edited by

Alfonso Pompella Department of Experimental Pathology, University of Pisa Medical School, Pisa, Italy

Gábor Bánhegyi Department of Medical Chemistry, Semmelweis University, Budapest, Hungary

Maria Wellman-Rousseau Faculté de Pharmacie, Université H. Poincaré, Nancy, France

IOS

Press

HI Ohmsha

Amsterdam • Berlin • Oxford • Tokyo • Washington, DC Published in cooperation with NATO Scientific Affairs Division

Proceedings of the NATO Advanced Research Workshop on Thiol Metabolism and Redox Regulation of Cellular Functions 10-13 April, 2002 Pisa, Italy © 2002, IOS Press All rights reserved. No part of this book may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, without prior written permission from the publisher. ISBN 1 58603 282 8 (IOS Press) ISBN 4 274 90542 X C3045 (Ohmsha) Library of Congress Control Number: 2002111980

Publisher IOS Press Nieuwe Hemweg 6B 1013 BG Amsterdam Netherlands fax:+31 20 620 3419 e-mail: [email protected]

Distributor in the UK and Ireland IOS Press/Lavis Marketing 73 Lime Walk Headington Oxford OX3 7AD England fax:+44 1865 75 0079

Distributor in the USA and Canada IOS Press, Inc. 5795-G Burke Centre Parkway Burke, VA 22015 USA fax: +1 703 323 3668 e-mail: [email protected]

Distributor in Germany, Austria and Switzerland IOS Press/LSL.de Gerichtsweg 28 D-04103 Leipzig Germany fax: +49 341 995 4255

Distributor in Japan Ohmsha, Ltd. 3-1 Kanda Nishiki-cho Chiyoda-ku, Tokyo 101-8460 Japan fax:+81 3 3233 2426

LEGAL NOTICE The publisher is not responsible for the use which might be made of the following information. PRINTED IN THE NETHERLANDS

Foreword The purpose of an Advanced Research Workshop could, in my mind, not have been better met as was done in this one on "Thiol Metabolism and Redox Cell Regulation", organized by Prof. Alfonso Pompella and Gábor Bánhegyi and their colleagues, held in Pisa, Italy, April 10-13, 2002. The topic has notable history, with thiol groups in proteins and with the low-molecular-mass compound, glutathione, being in research focus throughout the last century. However, very recent developments shedding light on the dynamics of cell regulation and function warrant the subtitle of the workshop: "New Evidence, Insight, and Speculation". This book contains state-of-the-art research in this developing field. It is hoped that the flavor of the vivid workshop is captured in this book for a general readership. The organizers are to be congratulated on their efforts, putting together an international group of experts working on oxidative signaling, glutathione and its metabolism, glutathione transferases and the related transport systems, and in particular the new roles of gamma-glutamyltransferase and of protein glutathionylation.

Helmut Sies

This page intentionally left blank

Preface Interest in the importance of cellular and exogeneous thiols in biomedical sciences continues to grow, with molecules such GSH or N-acetylcysteine making major impacts on biological, pharmacological and clinical knowledge. A number of recent studies stimulated intense interest in the role played by thiols in a range of key cell functions which might be under redox control. Newly discovered functions of cellular thiols are rapidly changing our perspective on several important aspects of cellular homeostasis. Cellular thiols glutathione (GSH) in the first place were for decades considered as just a defense line against prooxidant agents, seen as "injuring" species altogether. In recent years, however, it has been increasingly recognized that a number of cell types normally produce low amounts of prooxidants, and it is clear now that redox biochemistry of GSH and other thiols including protein thiols - occupies a crucial position in such "basal" production of prooxidants. Low, basal levels of thiol oxidation can exert physiological roles within the cell, e.g. in transduction of extracellular proliferative/apoptotic signals and regulation of gene expression; the term "oxidant-mediated regulation" seems today a more accurate alternative to the previous definition of oxidant stress (Cotgreave et al. 1998). Alterations in the redox status of protein thiols can mediate a nontoxic, physiological role of free radicals and other prooxidants in modulation of the function of growth factor receptors, protein kinases and transcription factors. Also, the pathophysiological potential of reactions of Sthiolation and dethiolation in the modulation of several enzymatic activities has been established, and fascinating aspects are emerging in the field of biochemistry of protein synthesis, where thiols and thiol-related enzymes appear to be essential for the correct folding of newly syntesized polypeptides. Against this background, it was felt that a Workshop on the biological and pathophysiological implications of thiol redox biochemistry could timely address some of the major implications arising from the change in perspective in thiol redox pathophysiology, and that the confrontation of different viewpoints could help to elucidate the significance of recently described phenomena, and foster the discussion of the most tempting views and hypotheses. The organizational process of the meeting stemmed from a pluriannual cooperation between A. Pompella and A. Paolicchi from the Dept. of Experimental Pathology, University of Pisa, and M. Emdin and C. Passino from the National Research Council Institute of Clinical Physiology, Pisa, a connection that allowed the clinical confirmation of interesting suggestions from basic experimental work. The workshop was hosted by the City of Pisa: here the Medical School began its activity in the 15th century, and the Italian National Res. Council has recently established its largest Research Area - the two Institutions traditionally cooperating in the field of biomedical research. The Workshop took place from 10 to 13 April 2002, and involved over 40 scientists from 18 different countries. The present volume includes 32 chapters, which were written by over 100 contributors with the precise aim of highlighting the speculative implications of their experimental work in the field of cell redox regulation. Alfonso Pompella

Gábor Bánhegyi

Maria Wellman-Rousseau

Contributors GABOR BANHEGYI + Dept. of Medical Chemistry, Semmelweis University - P.O. Box 260, 1444 Budapest VIII, Hungary GRZEGORZ BARTOSZ + Dept. of Molecular Biophysics, University of Lodz - Banacha 12/16, 90-237 Lodz, Poland AALT BAST + University of Maastricht, Faculty of Medicine, Dept. of Pharmacology and Toxicology, P.O. Box 616 - 6200 MD Maastricht, Netherlands ANGIOLO BENEDETTI + Dept. of Pathophysiology and Experimental Medicine, University of Siena - Via Aldo Moro, 53100 Siena, Italy HANS K. BIESALSKI + Dept. of Biological Chemistry & Nutrition, University of Hohenheim - Fruhwirthstrasse 12, D-70593 Stuttgart, Germany REGINA BRIGELIUS-FLOHE + Dept. Vitamins and Atherosclerosis, German Inst. of Human Nutrition - Arthur-Scheunert-Alice 114-116, D-14458 Bergholz-Rehbruecke, Germany BRIAN COLES + Univ. of Arkansas for Medical Sciences, Natl. Center for Toxicological Research, Div. of Molecular Epidemiology, HFT100, 3900 NCTR Road - Jefferson, AR 72079-9502 - USA IAN COTGREAVE + Div. of Biochemical Toxicology, Inst. of Environmental Medicine, Karolinska Institute - Box 210, S-17177 Stockholm, Sweden MARIO COMPORTI + Dept. of Pathophysiology and Experimental Medicine, University of Siena - Via Aldo Moro, 53100 Siena, Italy MIKLóS CSALA + Semmelweis University, Dept. of Medical Chemistry - P.O. Box 260, 1444 Budapest VIII, Hungary PETER CSERMELY + Semmelweis University, Dept. of Medical Chemistry - P.O. Box 260, 1444 Budapest VIII, Hungary SANDRINE DAUBEUF + "Thiols et fonctions cellulaires", Faculté de Pharmacie, Université H.Poincaré - 30, rue Lionnois - BP 403, 54001 Nancy, France MARC Diederich + RSL/Centre Universitaire de Luxembourg, 162A Avenue de la Faiancerie, L-1511 Luxembourg, Luxembourg MICHELE EMDIN + Inst. of Clinical Physiology, National Research Council (CNR), via Giuseppe Moruzzi 1- 56124, Pisa, Italy MILICA ENOIU + Dept. of Biochemistry, Faculty of Pharmacy, University of Medicine and Pharmacy "Carol Davila", Bucarest, Romania LEOPOLD FLOHE + Dept. of Biochemistry, Technical University of Braunschweig Mascheroder Weg, 1, D-38124 Braunschweig, Germany HENRY J. FORMAN + Dept. of Environmental Health Sciences, School of Public Health, Univ. of Alabama at Birmingham, 1530 3rd Avenue South, RPHB 317 - Birmingham, AL 35294-0022 - USA MARIE-MADELEINE GALTEAU + "Thiols et fonctions cellulaires", Faculte de Pharmacie, Université H.Poincaré - 30, rue Lionnois - BP 403, 54001 Nancy, France

HELEN GRIFFITHS + Pharmaceutical Birmingham B4 7ET, U.K.

Sciences, Aston University - Aston Triangle,

PHILIP J. HOGG + Centre for Thrombosis and Vascular Research, Univ. of New South Wales - Room 414C, Wallace Wurth Building, Gate 9, High Street - Sydney 2052, Australia LÁSZLÓ HOMOLYA + Membrane Research Group, Hungarian Academy of SciencesDiszegi u 24, 1113 Budapest, Hungary REBECCA P. HUGHEY + 933 Scaife Hall, Dept. Medicine - Renal, University of Pittsburgh School of Medicine - 3550 Terrace Street - Pittsburgh, PA 15213 - USA NILS ERIK HUSEBY + Dept. of Medical Biochemistry (1MB), Faculty of Medicine, University of Troms0, Norway DEAN P. JONES + Dept. of Biochemistry, Emory University School of Medicine - 4172 Rollins Research Center - 1510 Clifton Road, Atlanta, GA 30322 - USA MARTIN JOYCE-BRADY + The Pulmonary Center, Boston Univ. School of Medicine - 715 Albany Street, R304, Boston, MA 02119 - USA PIERRE LEROY + "Thiols et fonctions cellulaires", Faculte de Pharmacie, Université H.Poincaré - 30, rue Lionnois - BP 403, 54001 Nancy, France MICHAEL W. LIEBERMAN + Dept. of Pathology, Baylor College of Medicine - One Baylor Plaza, Houston, TX 77030 - USA JOE LUNEC + Div. of Chemical Pathology, Robert Kilpatrick Building, Clinical Sciences Level 0 - PO Box 65, LRINHS Trust - Leicester LE2 7LX, U.K. EMILIA MAELLARO + Dept. of Pathophysiology and Experimental Medicine, University of Siena - Via Aldo Moro, 53100 Siena, Italy JÓSZEF MANDL + Dept. of Medical Chemistry, Semmelweis University - P.O. Box 260, 1444 Budapest VIII, Hungary UMBERTO MURA + Dept. of Biochemistry, Faculty of Life Sciences, University of Pisa, Italy GÁBOR NARDAI + Inst. of Medical Chemistry, Semmelweis University, P.O. Box 260, 1444 Budapest VIII, Hungary GERHARD NöHAMMER + Institute of Molecular Biology, Biochemistry and Microbiology, Karl-Franzens-University - Heinrichstrasse 31 A, A-8010 Graz, Austria ALDO PAOLICCHI + Department of Experimental Pathology, University of Pisa Medical School, Via Roma 55 - 56126 Pisa, Italy CLAUDIO PASSING + Inst. of Clinical Physiology, National Research Council (CNR), via Giuseppe Moruzzi 1- 56124, Pisa, Italy ALFONSO POMPELLA + Department of Experimental Pathology, University of Pisa Medical School, Via Roma 55 - 56126 Pisa, Italy HELMUT SIES + Inst. fur Physiologische Chemie I, Heinrich-Heine-Universitat Düsseldorf - Postfach 101007, D-40001-Dusseldorf, Germany AVISHAY-ABRAHAM STARK + Dept. of Biochemistry, Tel Aviv University - Ramat Aviv 69978, Tel Aviv, Israel OREN TIROSH + Institute of Biochemistry, Food Science and Nutrition, The Hebrew University of Jerusalem, P.O. Box 12, Rehovot 76100, Israel

MARTINE TORRES + Childrens Hospital Los Angeles Research Institute, MS 57 - 4650 Sunset Blvd, Los Angeles CA 90027 - USA ATHANASE VISVIKIS + "Thiols et fonctions cellulaires", Faculté de Pharmacie, Universite H.Poincare - 30, rue Lionnois - BP 403, 54001 Nancy, France WALTER H. WATSON + Dept. of Biochemistry, Emory University School of Medicine 4172 Rollins Research Center - 1510 Clifton Road, Atlanta, GA 30322 - USA MARIA WELLMAN-ROUSSEAU + "Thiols et fonctions cellulaires", Faculte de Pharmacie, Universite H.Poincare - 30, rue Lionnois - BP 403, 54001 Nancy, France LECH WOJTCZAK + Nencki Institute of Experimental Biology - Pasteura 3, PL-02-093 Warsaw, Poland

Contents Foreword, Helmut Sies Preface Contributors

v vii viii

Oxidative Signaling and Glutathione Synthesis, H.J. Forman and DA. Dickinson 1 Cell Survival and Changes in Gene Expression in Cells Unable to Synthesize Glutathione, E. Rojas, Z.-Z. Shi, M. Valverde, R.S. Paules, G.M. Habib and M.W. Lieberman 14 Role of Glutathione in the Regulation of Liver Metabolism, J. Mandl and G. Bánhegyi 22 Glutathione Transport in the Endo/Sarcoplasmic Reticulum, M. Csala, R. Fulceri, J. Mandl, A. Benedetti and G. Bánhegyi 29 Role of Ascorbate in Oxidative Protein Folding, G. Bánhegyi, M. Csala, A. Benedetti and J. Mandl 38 Cytophotometric Investigations on Oscillating Thiol-Disulfide Equilibria and Oxidized Protein Sulfur, G. Nöhammer 48 Protection by Pantothenic Acid against Apoptosis and Cell Damage by Oxygen Free Radicals - The Role of Glutathione, L. Wojtczak and V.S. Slyshenkov 61 Thiols as Major Determinants of the Total Antioxidant Capacity, A. Balcerczyk, A. Grzelak, A. Janaszewska, W. Jakubowski, S. Koziol, M. Marszalek, B. Rychlik, M. Soszynski, T. Bilinski and G. Bartosz 15 Enzymes of the Thiol-dependent Hydroperoxide Metabolism in Pathogens as Potential Drug Targets, H. Budde and L. Flohe 85 Is there a Role of Glutathione Peroxidases in Signaling and Differentiation? R. Brigelius-Flohe and L. Flohe 96 Multidrug Resistance-associated Proteins: Export Pumps for Conjugates with Glutathione, Glucuronate or Sulfate, L. Homolya, A. Váradi and B. Sarkadi 107 Detoxification of Electrophilic Compounds by Glutathione S-Transferase Catalysis: Determinants of Individual Response to Chemical Carcinogens and Chemotherapeutic Drugs? B.F. Coles and F.F. Kadlubar 119 Transcriptional Regulation of Glutathione S-Transferase Pl-1 in Human Leukemia, A. Duvoix, M. Schmitz, M. Schnekenburger, M. Dicato, F. Morceau, M.-M. Galteau and M. Diederich 138 Mechanism of γ-Glutamyltranspeptidase Folding and Activation in the Endoplasmic Reticulum, R.P. Hughey 146 The Role of γ-Glutamyl Transpeptidase in the Biosynthesis of Glutathione, A.-A. Stark, N., Porat, G. Volohonsky, A. Komlosh, E. Bluvshtein, C. Tubi and P. Steinberg 160 Role of γ-Glutamyltransferase in the Homeostasis of Glutathione during Oxidative and Nitrosative Stress, N.-E. Huseby, N. Asare, S. Wetting, I.M. Mikkelsen, B. Mortensen and M. Wellman 172 The Importance of gamma-Glutamyl Transferase in Lung Glutathione Homeostasis and Antioxidant Defense, M. Joyce-Brady, Y. Liu, R.E. Marc and J.-C. Jean 182 The Role of gamma-Glutamyltranspeptidase in the Metabolism and Cytotoxicity of 4-Hydroxynonenal-Glutathione Conjugate: Evidence and Hypothesis, M. Enoiu, R. Herber, P. Leroy and M. Wellman 197

γ-Glutamyltransferase-Dependent Prooxidant Reactions: a Factor in Multiple Processes, S. Dominici, A. Paolicchi, E. Lorenzini, E. Maellaro, M. Comporti, L. Fieri, G. Minotti and A. Pompella Serum gamma-Glutamyl Transpeptidase: a Prognostic Marker in Cardiovascular Diseases, M. Emdin, C. Passino, A. Pompella and A. Paolicchi Lipoic Acid: a Multifunctional Antioxidant, A. Bast and G.R.M.M. Haenen Is Glutathione an Important Neuroprotective Effector Molecule against Amyloid Beta Toxicity? V.S. Barber and H.R. Griffiths Antioxidants in Cancer Therapy: is there a Rationale to Recommend Antioxidants during Cancer Therapy? H.K. Biesalski and J. Frank Disulfide Exchange in CD4, L.J. Matthias, P.T. W. Yam, X.-M. Jiang and P.J. Hogg Redox Regulation in Protein Folding and Chaperone Function, P. Csermely, G. Nardai and Cs. Soti Reduction of the Endoplasmic Reticulum Accompanies the Oxidative Damage of Diabetes Mellitus, G. Nardai, T. Korcsmdros and P. Csermely Analytical Developments in the Assay of Intra- and Extracellular GSH Homeostasis, LA. Cotgreave Signalling Potential and Protein Modifying Ability of Physiological Thiols, U. Mura M. Cappiello, P.G. Vilardo, I. Cecconi, M. Dal Monte and A. Del Cor so Redox Signaling and the Map Kinase Pathways, M. Torres and H.J. Forman Redox Regulation of Mitochondrial Permeability Transition: Contrasting Effects of Lipoic Acid and its Positively Charged Analog LA-Plus, O. Tirosh, S. Shilo, A. Aronis and C.K. Sen Redox State of Glutathione and Thioredoxin in Differentiation and Apoptosis, W.H. Watson, Y. Chen and D.P. Jones Redox Regulation of DNA Repair, J. Lunec, M.S. Cooke and M.D. Evans Author Index

209 223 230 238 252 265 273 281 290 299 306

317 328 338 349

Veniet tempus quo posteri nostri tam aperta nos nescisse mirentur (Seneca) A time will come when our descendants will wonder how could we ignore such overt matters...

This page intentionally left blank

Thiol Metabolism and Redox Regulation of Cellular Functions A. Pompella et al. (Eds.) IOS Press. 2002

Oxidative Signaling and Glutathione Synthesis Henry Jay FORMAN and Dale A. DICKINSON Department of Environmental Health Sciences, School of Public Health, University of Alabama at Birmingham, 1530 3*° Avenue S, RPHB-317, Birmingham, AL 35294-0022. Tel: 205-975-8949 Fax: 205-975-6341. E-mail: hformanCcb.uab.edu

1. Prologue The major endogenous small molecular weight thiol, glutathione (GSH, Y-ghitamyl-cysteinylglycine) has roles in cellular protection against oxidants and xenobiotics, and in signal transduction. In antioxidant defense, the major reaction of GSH is reduction of hydroperoxides by glutathione peroxidases (GSHPx) and at least one peroxiredoxin, which yields glutathione disulfide (GSSG). GSHPx also catalyzes reduction of peroxynitrite to nitrite, which also yields GSSG. GSSG is usually rapidly reduced by glutathione reductase but it and other glutathioneconjugates may also be excreted from cells. In physiological redox signaling, GSH participates through both the removal of HaCh, which has the properties of a second messenger, and the reversal of thiolate (S") oxidation. For example, protein tyrosine phosphatases contain thiolate residues in their active sites that are converted to a catalytically inactive sulfenic acid through reaction with H2O2 during cell signaling but are then reduced by GSH to restore the active form. In contrast, during oxidative stress, GSSG is elevated through the GSHPx reaction and consequently, formation of glutathione-mixed disulfides results in a less specific type of signaling process. GSH synthesis occurs in two enzymatically catalyzed steps: The first, catalyzed by glutamate cysteine ligase (GCL), which produces y-glutamyl cysteine, is generally considered to be rate limiting. Glutathione synthase adds glycine to complete the synthesis of GSH. GCL is composed of two subunits and is regulated at the transcriptional, translational, and posttranslational levels. Although the catalytic subunit (GCLC) does not require the modulatory subunit (GCLM) for catalytic activity, the kinetics under physiological conditions are largely controlled by GCLM. Transcription of both GCL subunits is induced by a wide variety of agents. For oxidative induction, putative cis elements regulating transcription of Gel mRNAs include TRE (the AP-1 binding site), ARE (EpRE), and KB (absent from Gclm). Investigation of transcriptional regulation has been intense but has yielded a complex picture in which the pathways for up-regulation of the subunits appear to be independent and vary with inducing agent and cell type. As examples, Gclc is induced through an ARE element by pnaphthoflavone in human hepatoma cells, while AP-1 activation through the INK pathway appears responsible for Gclc induction by the lipid peroxidation product, 4-hydroxynonenal in human bronchial epithelial cells. Thus, GSH is intimately entwined in redox signaling as both a participating factor and in its own synthesis being regulated through redox signaling.

H.J. Forman and D.A. Dickinson / Oxidative Signaling and Glutathione Synthesis

2. GSH maintains Normal Cellular Physiology Thiol-containing compounds have an essential role in many biochemical and pharmacological reactions due to the ease with each they are oxidized, and the rapidity with which they can be regenerated. This is especially evident for thioredoxin (Trx) and glutathione (GSH), two of the major small molecular weight thiol-containing compounds synthesized de novo in mammalian cells that participate in those functions. As the predominant non-protein sulfhydryl in cells, GSH plays several important roles. It has long been established that the thiol moiety of GSH is important in antioxidant defense, xenobiotic and eicosanoid metabolism, and regulation of the cell cycle and gene expression (for reviews see [1-5]). Although GSH does not react directly with hydroperoxides, its use as a substrate for glutathione peroxidase (GSHPx) has been recognized for almost forty years as the predominant mechanism for reduction of H2O? and lipid hydroperoxides [6]: H2O2 + 2 GSH — Glutathione peroxidase -* 2 H2O + GSSG ROOH + 2 GSH — Glutathione peroxidase -» 2ROH + GSSG

<2>

where ROOH and ROH are a hydroperoxide and its corresponding alcohol. The GSHPxs are a family of selenoproteins that vary in their hydroperoxide substrate specificity [7]. More recently, a family of proteins, now called peroxiredoxins, has been recognized as catalyzing the reduction of H2O2 by GSH and/or other thiols, but with cysteine, in its thiolate (S~) form in their active sites rather than selenium. The reaction sequence is: H2O2 + Prx-S~ -» OH~ + Prx-SOH

<3>

Prx-SOH + RSH -* Prx-SSR + H2O

<4>

Prx-SSR + RSH — Prx-S~ + RSSR

<5>

where Prx-SOH represents the sulfenic acid intermediate and RSH represents a thiol. For most of the peroxiredoxins, the thiol is Trx and the disulfide formed is intramolecular. Nonetheless, at least one of the peroxiredoxins seems to prefer GSH as its thiol substrate forming GSSG in the reaction. Glutathione disulfide, which should not be referred to as oxidized glutathione, is reduced to GSH by NADPH through the glutathione reductase (GR) reaction: GSSG + NADPH + H+ — Glutathione reductase — NADP* + 2 GSH

<6>

while NADPH is maintained predominantly by the pentose phosphate shunt. GSSG is normally maintained as less than 1% of total glutathione. Increases in GSSG during oxidative stress are generally transient as reduction by GR is relatively rapid. Nonetheless, GSSG can exchange with protein sulfhydryls to produce protein-glutathione mixed disulfides [8]: GSSG + Protein-SH *» Protein-SSG + GSH

<7>

H.J. Forman and D.A. Dickinson / Oxidative Signaling and Glutathione Synthesis

3

The mixed disulfides (Protein-SSG) have a longer half-life than GSSG, probably due to protein folding, and a significant basal level is found in cells [9]. This exchange reaction provides an important mechanism for the action of GSH in cell signaling (see below). An ATP-dependent transport mechanism is also responsible for decreasing GSSG through export [10]. Although these aspects of GSH thiol chemistry are generally well established, the literature abounds with cases in which the details are ignored in the interpretation of experimental results. GSH forms conjugates with a great variety of electrophilic compounds nonenzymatically, when the electrophile is very reactive, or more often through the action of glutathione S-transferases (GST). Conjugation with GSH is an essential aspect of both xenobiotic and normal physiological metabolism (for reviews see [11, 12]). Formation of conjugates can result in depletion of GSH and has been used as a tool to study the role of GSH in antioxidant defense. A caveat is that the use of a strong electrophile that does not require catalysis will react with protein thiols as well, producing non-specific responses. While GSH does not react nonenzymatically with H2O2, another role for GSH in antioxidant defense that depends upon its ability to react with carbon centered radicals (R) has been proposed by Winterbourn [13]. In this "free radical sink" hypothesis, GSH acts in concert with superoxide dismutase to prevent oxidative damage: R' + GSH -^ RH + GS'

<8>

GS* + GS" -» GSSG*"

<9>

GSSG'" + O2 -* GSSG + O2'~

<10>

2O 2 "~ + 2H+

<11>

—Superoxide dismutase-* H2O2 + O2

The critical role of GSH in protecting cells should be apparent from the multitude of reactions through which it removes potentially harmful molecules from cells. While restoration of GSH from GSSG can be easily accomplished, depletion through conjugation or loss by excretion of GSSG demands replenishment. Although a few cells can take up GSH, de novo synthesis is the predominant pathway for this restoration. The signaling for de novo synthesis will be discussed later, and is perhaps the only aspect of GSH-related redox signaling significantly addressed in the literature to date. An intriguing and little-addressed question remains: How do changes in the content of GSH effect redox signaling in general?

3. Changes in GSH Content effects Redox Signaling GSH has one cysteine and is found in the millimolar range in most cells. Changes in GSH content are typical in the response of a cell to a stress. Reactions that protect the cell by removing or altering deleterious compounds consume GSH. This temporary depletion must then be reversed through either enzymatic reduction of a disulfide or by de novo synthesis, to restore baseline levels of GSH. These changes in GSH content and metabolism can have profound effects on signaling pathways, probably through alteration of the redox state. For example, thiol-containing compounds such as GSH are central in many biochemical and

4

H.J. Forman and D.A. Dickinson / Oxidative Signaling and Glulathione Synthesis

pharmacological reactions because disulfide bonds have important roles in determining the tertiary structure of proteins, and, in many enzymes the cysteine moiety is involved in catalysis. PubMed lists 2311 papers in a search for "redox state" as of the date this is being written. The question of how to define the cellular redox state has recently been examined by Buettner and colleagues [14], who determined that the ratio (2 GSH:GSSG) is the best definition of that term. But how does this ratio translate into redox signaling? Disulfide exchange between GSSG and protein thiols is one obvious hypothesis, but as protein thiols are ubiquitous, the specificity required for signaling is absent in this scenario. Under sublethal oxidative stress where I^Oi concentration is dramatically elevated, an increase in GSSG will occur through the GSHPx reaction (reaction <1>). As a consequence, formation of protein mixed disulfides will also increase. As many proteins involved in signaling have critical thiols, such as receptors and several transcription factors, the formation of mixed disulfides with their cysteine residues can lead to alteration of activity. Thus, the number of reported functional alterations by oxidative stress that are influenced by glutathione metabolism continues to grow at a rapid pace. Nonetheless, while these alterations are an important aspect of pathophysiology and probably trigger adaptive responses when the stress is not lethal, the signaling is very broad compared to what is observed in response to physiologically generated HbO2. The signaling leading to the adaptive response elicited in response to sublethal oxidative stresses is an area of research that remains virtually unstudied. Under normal physiological conditions, GSH probably functions in signaling in a less direct fashion; in being a determinant of the rate of H2O2 removal by glutathione peroxidase and being a reductant of reactions of HaCh, which is the more likely direct agent of redox signaling. The specific reactions in signaling by H2O2 involve a different aspect of thiol chemistry than its reaction in the glutathione peroxidase reaction or non-enzymatic metal catalyzed oxidations. Sulfinic (-SOiH) or sulfonic (-SOsH) acids, which are not easily reduced, are formed during metal-catalyzed reactions of fyCh or peroxynitrite with a thiol. HiCh is unreactive with thiols at physiological pH in the absence of catalysis. But, thiolate anion (S~) found in some proteins can react with H2C>2 produced during signaling to form a sulfenic acid as in reaction <3>. As cysteine normally has a pKa significantly higher than the physiological range, thiolates are not common in proteins, providing a specificity that would be lacking if any thiol were the target. Importantly, thiolates are found in the active sites of some proteins involved in signaling, including peroxiredoxins, protein tyrosine phosphatases (FTP) and the thioredoxin (Trx) family of proteins [15-18]. These proteins provide an unusually basic microenvironment in their actives sites in which cysteine dissociates to form the thiolate. The oxidation of a FTP thiolate resulting in loss of activity has been suggested by in vitro and indirect cellular experiments as a mechanism for H2O2 signaling [16, 19]. GSH could then have a direct role in signaling as the sulfenic acid can be easily reduced in two sequential reactions involving GSH (as in reactions <4> and <5>), restoring the FTP activity. Such restoration of activity is important as it allows the "turn-off' of a signaling pathway. Trx family proteins also have a thiolate that can form a sulfenic acid; however, a second conserved cysteine reacts with the sulfenic acid to form an intramolecular disulfide bridge. This intramolecular disulfide cannot be reduced by GSH and is instead reduced by NADPH in a reaction catalyzed by Trx reductase, a selenoprotein. Ichijo and coworkers have shown that activation of ASK1, an upstream kinase in the JNK pathway, is regulated through reversible Trx oxidation [20]. There are several important issues to resolve, however. Reported rates of thiolate oxidation vary markedly and seem to be low enough to question their involvement in signaling. One potential resolution of this issue will require determining the

H.J. Forman and D.A. Dickinson / Oxidative Signaling and Glutathione Synthesis

5

rate of HiOi or Ch reaction with the thiolate. It is also possible that a low rate would still allow signaling if the target protein is sufficiently close to site of HaCh production to allow a significantly high H^Oj concentration to be present near the thiolate. The finding that signaling proteins are in close proximity, often involving specialized scaffold proteins to hold them in place, has become increasingly frequent. Location may therefore be as important, if not more so, in redox signaling as in other signal transduction pathways.

4. Redox Signaling for GSH Biosynthesis We have seen how the content of GSH can relate to the redox state of the cell, with key roles in redox signaling. The content of GSH in the cell is a balance between depletion via use in protective reactions, and replacement, via either reduction of a disulfide or by de novo synthesis. The synthesis of GSH from its constituent amino acids occurs both constitutively to maintain basal levels for use in normal physiology and in maintaining the redox state of the cells, and is regulated in response to stress, especially stresses that alter the cellular redox state. Biosynthesis of GSH results from the concerted effort of two ATP-dependent enzymes: L-Glutamate + Cysteine + ATP —GCL-» y-L-glutamyl-L-cysteine + ADP + Pi

<12>

y-L-glutamyl-L-cysteine + Glycine + ATP — GS -» —» y-L-glutamyl-L-cysteinyl-glycine + ADP + Pj

<13>

This first enzyme, following IUBMB recommendations, should be called glutamate-cysteine ligase (GCL), but has been commonly referred to as v-glutamylcysteine synthetase (GCS), and other, less-common names. It is the rate-limiting step in de novo synthesis. Under nondenaturing conditions the heterodimeric GCL can be dissociated into a modulatory, or light, subunit (GCLM), and a catalytic, or heavy, subunit (GCLC) [21]. GCLC is 73 kDa in size, possesses all of the catalytic activity of GCL, and is the site of GSH feedback inhibition. The lighter, 31 kDa GCLM subunit exhibits a modulatory or regulatory effect on the GCLC subunit when associated with it; it has no known catalytic function as a monomer. The association of both subunits is probably essential for GSH biosynthesis under normal physiological concentrations of glutamate and GSH, based on results from experiments performed in vitro using purified rat [22] or recombinant human enzymes [23, 24], and on observations made in vivo from transgenic mice [25], all of which suggest that the major effect of the light subunit in vivo is on elevating the Kj for GSH such that it decreases negative feedback inhibition. The kinetics of this association results in the interesting and generally ignored fact that acute depletion of GSH can lead to a short-term increase in GSH synthesis because, to some extent, a decrease in GSH will cause a transient increase in the activity of pre-existing GCL by decreasing the feedback inhibition by GSH, resulting in a short-term increase in GSH synthesis [26]. The activity of the GCL holoenzyme can further be regulated either positively or negatively by S-nitrosation [24], phosphorylation [27] and oxidation [28], although increased GCL activity in most cases involves a transcriptional component leading to increased production. The second enzyme required for de novo GSH biosynthesis, according to IUBMB nomenclature, is properly named glutathione synthase (GS), although it is more often referred to as glutathione synthetase. GS functions as a homodimer of 118 kDa, and is responsible for

6

H.J. Forman and D.A. Dickinson I Oxidative Signaling and Glutathione Synthesis

the addition of glycine to the y-glutamylcysteine created by GCL to form GSH. While less is known about the regulation of GS activity compared with GCL, certain clinical phenotypes have been found to result from the inheritance of missense mutations in Gs. Most of these mutations decreased either the Km for glycine and/or the Vma* value, or caused a decrease in the stability of the enzyme [29]. Pathologically, a systemic decrease in GSH content is a standard clinical phenotype associated with HIV-infected persons [30, 31], although significant controversy remains concerning the mechanisms of GSH depletion. Recently, our lab has demonstrated that in the livers and erythrocytes of Tat+ transgenic mice this decrease in GSH results from a decrease in Gclm mRNA and protein content, and a significant reduction in the activity of GS [25]. Specifically, downregulation of Gclm in Tat+ mice was associated with an increased sensitivity of GCL to feedback inhibition by GSH, which is likely to be partially responsible for the observed decreased level of GSH. Furthermore, GS activity was also decreased, and was found to linearly correlate with the GSH content. We propose the HIV Tat protein causes a perturbation in the intracellular GSH level, leaving the cells more vulnerable to damage by oxidants and xenobiotics. Increased drug toxicity and oxidative damage is often found in HIV infected individuals and suggests the importance of GSH in disease progression; and, GSH content may even predict the survival of HIV-seropositive individuals [32,33]. The two key enzymes responsible for GSH biosynthesis are encoded by single-copy genes in the haploid human genome; two for GCL and one for GS. The cDNAs for both Gel genes have been cloned and sequenced [34, 35], and the 5' untranslated regions of both have been cloned and sequenced, and analyzed for potential regulatory elements [36-38]. The cDNA for Gs has been cloned and sequenced [39]. Unfortunately, as the 5' untranslated region of Gs has not been cloned, and is not well represented in the expressed sequence tag (EST) databases, potential regulatory elements for this gene remain unknown. An increase in GSH biosynthesis in response to compounds that form glutathioneconjugates, or those that generate reactive oxygen species, is often correlated with increased transcription of the Gel genes (see [40] for review). Sequence analysis and experimental manipulations employing reporter constructs of the 5' untranslated regions for the human Gel genes have revealed several putative enhancer elements that could mediate, either alone or in combination, an increase in transcription in response to the binding of transcription factors, whose activity has been increased in response to a stimulus signaled by the presence of a compound. Although architecturally quite different, both Gel promoters contain many of the same potential as-acting elements, including consensus recognition sites for binding of Sp-1, activator protein-1 (TRE), TRE-like, and the electrophile response element (EpRE, sometimes called the antioxidant response element, ARE) binding complexes [36, 37]. Only the Gclc promoter contains a nuclear factor kappa B (KB) element [38]. Despite these similarities, the Gel genes can be differentially regulated [41], although other confounding factors including subunit availability also need to be addressed. Of the above-mentioned enhancer elements, those that have received the most attention with respect to redox signaling have been TRE or TRE-like elements, and EpRE elements. The role of these elements in mediating Gel transcription in response to various stimuli has been reviewed previously [42]. It appears, at least for oxidative and xenobiotic stresses, that both EpRE and TRE sites are involved in Gel induction, and that differences in the expression of signaling components and metabolism among cell types causes signaling for activation of the corresponding transcription factors to vary even with the same compound. These apparent complexities and peculiarities of GSH regulation could be a manifestation of overlapping pathways used by the cell to ensure that adequate amounts of GSH are available for defense, and are as such, reflective of normal

H.J. Forman and D.A. Dickinson / Oxidative Signaling and Glutathione Synthesis

1

cellular signaling pathways, which are themselves inherently cell-type specific. Perhaps investigations into GSH regulation, as a model, may yield broader advances in cell-type specific signaling. Many different conditions are known to change intracellular GSH content. These include the presence of heavy metals [43], high glucose concentrations [44] and heat shock [45]. Exposure to reactive oxygen and nitrogen species including FbOa [46] and nitric oxide [47], or to compounds that can generate reactive species including 2,3-dimethoxy-l,4~ naphthoquinone [48], menadione [46, 48], tertiary-butylhydroquinone [49, 50], pyrrolidine dithiocarbamate [51] and p-naphthoflavone [37], and other reactive biological products such as 15-deoxy-A(12,14)-prostaglandin J2 [52], low density lipoproteins [53], and 4-hydroxy-2nonenal [54, 55] can increase the content of GSH by increasing the rate of GSH synthesis. Increased synthesis of GCL subunits through a combination of increased transcription and mRNA stability is the principle mechanism used to increase the rate of de novo synthesis, although as discussed above, removal of feedback inhibition via a temporary decrease in GSH content can lead to increased activity of pre-existing GCL [54, 56, 57]. The mechanisms by which compounds change GSH content have been investigated, most often at the level of Gel gene expression; relatively little has been reported on the roles of mRNA stability, increased rate of translation, and modifications that lead to increased holoenzyme activity. Perhaps the most important insight gained from reviewing expression studies is the differences among inducers upon Gel gene induction, GCL subunit content, and when studied, the differences in signaling pathways. One particularly well-studied inducer of GSH biosynthesis, and an intriguing example, is the lipid peroxidation product, 4-hydroxynonenal (4HNE).

5. 4HNE Signaling and GSH Metabolism The interaction of reactive oxygen species with the n-6 polyunsaturated lipids in cellular membranes, a normal part of physiology that is dramatically and damagingly increased during inflammation and exposure to air pollutants such as nitrogen dioxide and ozone, can lead to the formation the 4HNE [58-60]. This a,p-unsaturated aldehyde is relatively stable in vivo, and because of this has been proposed as being one of the key mediators of the damage resulting from exposure to reactive oxygen and nitrogen species [61]. 4HNE is removed from many cell types by reactions with GSH, catalyzed by the glutathione S-transferase (GST) subclasses that have relative specificity for alkenals (GSTA4-4 and GST5.8), the expression of which are regulated by their substrates, including 4HNE [62, 63]. Another major pathway for removal of 4HNE is its conversion to 4-hydroxynonenol by an aldehyde reductase [64] that is also inducible by 4HNE [65], while a third pathway is its oxidation by an aldehyde dehydrogenase to 4-hydroxy-2-nonenoic acid [66]. The relative contributions of each of these pathways in the removal of 4HNE has been reported for aortic endothelial cells [67] and isolated perfused rat heart [68]. Exposure of a rat alveolar epithelial cell line (L2 cells) to 4HNE caused an increase in GSH biosynthesis [54]. Similar results were obtained with a normal human bronchial epithelial cell line, HBE1 (unpublished data). Separately, 4HNE has been demonstrated to activate the c-Jun N-terminal kinase (JNK) signaling pathway [69-71]. JNK activation can in turn activate the activator protein-1 (AP-1) transcription factor complex through phosphorylation of Jun family members, c-Jun and JunB. There are many possible AP-1 transcription factor complexes, which result from dimerization of Jun family proteins as either homodimers, or as heterodimers with other Jun family

8

H.J. Forman and DA. Dickinson / Oxidative Signaling and Glutathione Synthesis

members, Fos family members, or other proteins, such as ATF-2. These various AP-1 dimers bind to TRE elements and TRE-like elements. Jun family members can also pair with Nrf2 and small Maf proteins to form EpRE-binding complexes [72]. Much less is understood concerning the composition of EpRE-binding complexes, but it appears that specificity, similar to AP-1 complexes, is conferred by the members of each EpRE-binding complex. Recent, unpublished data from our lab have demonstrated that exposure of HBE1 cells to physiologically relevant levels of 4HNE caused a dose- and time-dependent increase in the intracellular content of GSH, which was closely correlated with an increase in the content of both GCL subunits. Not surprisingly, this increase in the subunit content was correlated with, and kinetically preceded by, an increase in the steady-state level of both Gel mRNA species, which also occurred in both dose- and time-dependent fashions. Based on the putative cisacting elements in the Gel promoters identified as being likely to mediate transcription in response to oxidative stress, we used the electrophoretic mobility shift assay (EMSA) to investigate the activation of AP-1, EpRE and NF-icB transcription factor complexes in response to 4HNE exposure. We showed that only the AP-1 binding complex was activated in response to 4HNE exposure, whereas EpRE and NF-KB binding complexes showed no change in DNA binding activity. Super-shift analyses using antibodies specific to various potential AP-1 complex members demonstrated the presence of Jun family proteins. Furthermore, the content of phosphorylated c-Jun, a common component of activated AP-1 transcription factor complexes, also increased with 4HNE exposure. These collective results suggested that in HBE1 cells 4HNE was signaling for increased GSH biosynthesis via activation of the INK signaling pathway. Results from previous studies have suggested roles for TRE and EpRE elements in mediating Gel transcription, with signaling occurring through the MAPK pathways. So, the role of MAPK pathways in mediating the Gel response to 4HNE was examined next in HBE1 cells. Inhibition of the p38MAPK pathway with the pharmacological inhibitor SB202190 had no effect on either AP-1 binding activity or on the steady-state level of Gel mRNAs. Similarly, inhibition of the ERK signaling pathway with the pharmacological inhibitor PD98059 had no effect on AP-1 binding activity, or on steady-state Gel mRNA content. Using deductive reasoning, these results suggest that either the INK pathway mediates the effects of 4HNE, or that non-MAPK pathways are responsible. This question can finally be addressed by the use of either peptide-based, membrane-permeable inhibitors of INK signaling [73] or with the use of the new pharmacological inhibitor, SP600125 [74]. Results from these experiments will reveal the role of MAPK signaling, and INK signaling specifically, in mediating GSH synthesis in response to 4HNE. These results are eagerly anticipated, and will definitively show the mechanism of 4HNE action in GSH synthesis in normal human bronchial epithelial cells. Similar work done by us using 4HNE in rat L2 cells showed no effect of inhibiting the p38 pathway, and is similar to the result obtained with HBE1 cells. Inhibition of the ERK pathway, however, did effect the steady-state mRNA content of Gclc, but not Gclm. The results obtained from the L2 cells, when viewed in isolation, suggested to us partial involvement of both the INK pathway, by deductive reasoning, and the ERK pathway in Gclc expression, and for Gclm expression, suggested that the INK pathway has a major role in signaling [55]. Moreover, another interesting difference in the signaling for the two Gel genes in L2 cells by 4HNE was discovered even previous to this, with the demonstration that there existed a requirement for de novo protein synthesis for Gclm transcription but not for Gclc transcription, revealing that the signaling pathways are indeed different [54]. The aggregate knowledge from these three studies, conducted by the same group with the same inducer in similar cells types, vitally underscores the species-specific differences that exist between rat

H.J. Forman and D.A. Dickinson / Oxidative Signaling and Glutathione Synthesis

9

and human pathways controlling for GSH biosynthesis. We believe the results from these studies diminish the relevance of further use of rat models for studying aspects of GCL regulation if relevance to human physiology and pathophysiology are to be gained. This cautionary opinion, if considered, helps to clarify some of the apparent discrepancies that abound concerning signaling in GSH synthesis, given the broad range of cell types and species used in the various studies. Not surprisingly, other research illustrates the importance of not generalizing results obtained from one inducer to another. Using human hepatoma cells, Mulcahy and co-workers have demonstrated equivalent roles of both the p38 and ERK pathways in mediating the signaling for GSH production and Gel transcription in response to PDTC [75].

6. Epilogue GSH is a major component of the process for defense against the toxicity of xenobiotic compounds and oxidants to which exposure is an everyday occurrence. Normal metabolism requires constant and rapid replenishment of GSH, which is accomplished through both the reduction of GSSG and de novo synthesis. Determining which signaling pathways lead to alterations in GSH metabolism is critical for understanding the mechanisms of, and developing therapies for, environmental toxicants. We hypothesize that many environmental agents exert their deleterious effects by altering, either directly or indirectly, the cellular redox status through manipulation of the metabolism of thiols such as GSH. This was exemplified in the results with 4HNE, a reactive aldehyde produced in normal metabolism but which becomes markedly elevated during inflammation or in response to exposure to pollutants such as N(>2. Knowing how such agents alter these essential and ubiquitous biochemical pathways should facilitate the understanding of redox-mediated changes in other pathways and pathologies. Our aim here is point out that conclusions from any one investigation of the signaling for GSH synthesis can not usually be generalized, and that perturbations in any of step of thiol metabolism may have etiological roles in genetically, virally, and environmentally borne pathologies. We speculate that the importance of GSH in so many facets of cell biology requires multiple and redundant mechanisms to ensure production of this essential and abundant cellular constituent.

Acknowledgements This work was supported by grant ES05511 from the National Institutes of Health. We thank the workers in our lab that contributed to our work, particularly Drs. Rui-Ming Liu, Karen lies and Jinah Choi, and to the many people who have contributed to understanding the roles of thiols in biology, most of whose work could not be cited in such a brief commentary.

References 1. B. Ketterer, The role of nonenzymatic reactions of glutathione in xenobiotic metabolism, Drug Metab. Rev. 13 (1982) 161-187. 2. A. Meister, Selective modification of glutathione metabolism, Science 220 (1983) 472-477.

10

H.J. Forman and DA. Dickinson I Oxidative Signaling and Glutathione Synthesis

3. D.M. Ztegler, Role of reversible oxidation-reduction of enzyme thids-disulfides in metabolic regulation, Artnu. Rev. Biochem. 54 (1985) 305-329. 4. D.J. Reed and M.W. Fariss, Glutathione depletion and susceptibility, Pharmacd. Rev. 36 (1984) 25S-33S. 5. A.P. Arrigo, Gene expression and the thiol redox state, Free Radic. Bid. Med. 27 (1999) 936-944. 6. G. Cohen and P. Hochstein, Glutathione peroxidase: the primary agent for the elimination of hydrogen peroxide in erythrocytes, Biochemistry 2 (1963) 1420-1428. 7. R. Brigelius-Ftohe, Tissue-specific functions of individual glutathione peroxidases. Free Radic. Biol. Med. 27 (1999)951-965. 8. R. Brigelius, C. Muckel, T.P.M. Akerboom and H. Sies, Identification and quantitation of glutathione and its relationship to glutathione disulfide, Biochem. Pnarmacol. 32 (1983) 2529-2534. 9. G.A. Loeb, D.C. Skelton and H.J. Forman, Dependence of mixed disulfide formation in alveolar macrophages upon production of oxidized glutathione: effect of selenium depletion, Biochem. Pharmacd. 38 (1989) 31193121. 10. R. Sharma, S. Gupta, H. Ahmad, G.A.S. Ansari and Y.C. Awasthi, Stimulation of a human erythrocyte membrane ATPase by glutathione conjugates, Toxicol. Appl. Pharmacol. 104 (1990) 421-428. 11. R.C. Strange, P.W. Jones and A.A. Fryer, Glutathione S-transferase: genetics and rote in toxicology. Toxicol. Lett. 112-113 (2000) 357-363. 12. D.L Eaton and T.K. Bammter, Concise review of the glutathione S-transf erases and their significance to toxicology, Toxicol. Sci. 49 (1999) 156-164. 13. C.C. Winterboum, Superoxide as an intracellular radical sink. Free Radic. Bid. Med. 14 (1993) 85-90. 14. F.Q. Schafer and G.R. Buettner, Redox environment of the cell as viewed through the redox state of the glutathione disutfide/glirtathione couple, Free Radic. Biol. Med. 30 (2001) 1191-1212. 15. A. Holmgren, Thioredoxin structure and mechanism: conformational changes on oxidation of the active-site sulfhydryls to a disulfide, Structure 3 (1995) 239-243. 16. J.M. Denu and K.G. Tanner, Specific and reversible inactivation of protein tyrosine phosphatases by hydrogen peroxide: evidence for a sulfenic acid intermediate and implications for redox regulation. Biochemistry 37 (1998)5633-5642. 17. A. Claibome, J.I. Yeh, T.C. Maltett, J. Luba, E.J. Crane, 3rd, V. Charrier and D. Parsonage, Protein-sulfenic acids: diverse roles for an unlikely player in enzyme catalysis and redox regulation, Biochemistry 38 (1999) 15407-15416. 18.J.R. Kim, H.W. Yoon, K.S. Kwon, S.R. Lee and S.G. Rhee, Identification of proteins containing cysteine residues that are sensitive to oxidation by hydrogen peroxide at neutral pH, Ann. Biochem. 283 (2000) 214-221. 19. S.R. Lee, K.S. Kwon, S.R. Kim and S.G. Rhee, Reversible inactivation of protein-tyrosine phosphatase 1B in A431 cells stimulated with epidermal growth factor, J. Biol. Chem. 273 (1998) 15366-15372. 20. M. Saitoh, H. Nishitoh, M. Fuji!, K. Takeda, K. Tobiume, Y. Sawada, M. Kawabata. K. Miyazono and H. teWjo, Mammalian thioredoxin is a direct inhibitor of apoptosis signal-regulating kinase (ASK) 1, EMBO J. 17 (1998) 2596-2606. 21.G.F. Seelig and A. Meister, Glutathione biosynthesis: •y-glutamylcysteine synthetase from rat kidney, Methods Enzymol. 113 (1985) 379-390. 22. C.S. Huang, L.S. Chang, M.E. Anderson and A. Meister, Catalytic and regulatory properties of the heavy subunit of rat kidney gamma-glutamylcysteine synthetase, J. Bid. Chem. 268 (1993) 19675-19680. 23. Z. Tu and M.W. Anders, Expression and characterization of human glutamate-cysteine ligase, Arch. Biochem. Biophys. 354 (1998) 247-254. 24.0.W. Griffith, Biologic and pharmacologic regulation of mammalian glutathione synthesis, Free Radic. Bid. Med. 27(1999)922-935. 25. J. Choi, R.-M. Liu, R.K. Kundu, F. Sangiorgi, W. Wu, R. Maxson and H.J. Forman, Molecular mechanism d decreased glutathione content in human immunodeficiency virus type 1 Tat-transgenic mice, J. Bid. Chem. 275 (2000) 3693-3698.

H,J. Forman and D.A. Dickinson / Oxidative Signaling and Glutathione Synthesis

26. P.G. Richman and A. Meister, Regulation of gamma-glutamyl-cysteine synthetase by nonallosteric feedback inhibition by glutathione, J. Biol. Chem. 250 (1975) 1422-1426. 27. W.-M. Sun, Z.-Z. Huang and S.C. Lu, Regulation of y-glutamylcysteine synthetase by protein phosphorylation, Biochem. J. 320 (1996) 321-328. 28. T. Ochi, Hydrogen peroxide increases the activity of gamma-glutamylcysteine synthetase in cultured Chinese hamster V79 cells, Arch. Toxicol. 70 (1995) 96-103. 29. R. Njalsson, K. Carlsson, B. Olin, B. Cartsson, L Whitbread, G. Polekhina, M.W. Parker, S. Norgren, B. Mannervik, P.G, Board and A. Larsson, Patients with GS deficiency present with various clinical manifestations, including metabolic acidosis, 5-oxoprolinuria, progressive neurological dysfunction, and haemolytic anemia, Biochem. J. 349 (2000) 275-279. 30. H.P. Eck, H. Gmunder, M. Hartmann, D. Petzoldt, V. Daniel and W. Droge, Low concentrations of acid-soluble thiol (cysteine) in the blood plasma of HIV-1 -infected patients, Bioi. Chem. Hoppe-Seyler 370 (1989) 101-106. 31. R. Buhl, H.A. Jaffe, K.J. Holroyd, F.B. Wells, A. Mastrangeli, C. Saltini, A.M. Cantin and R.G. Crystal, Systemic glutathione deficiency in symptom-free HIV-seropositive individuals, Lancet 2 (1989) 1294-8. 32. LA. Herzenberg, S.C. De Rosa, J.G. Dubs, M. Roederer, M.T. Anderson, S.W. Ela and S.C. Deresinski, Glutathione deficiency is associated with impaired survival in HIV disease, Proc. Natl. Acad. Sci. U S A 94 (1997)1967-1972. 33. M. Marmor, P. Alcabes, S. Titus, K. Frenkel, K. Krasinski, A. Penn and R.W. Pero, Low serum thiol levels predict shorter times-to-death among HIV- infected injecting drug users, Aids 11 (1997) 1389-1393. 34. J.J. Gipp, C. Chang and R.T. Mulcahy, Cloning and nucleotide sequence of a full-length cDNA for human liver gamma-glutamylcysteine synthetase, Biochem. Biophys. Res. Comm. 185 (1992) 29-35. 35. J.J. Gipp, H.H. Bailey and R.T. Mulcahy, Cloning and sequencing of the cDNA for the light subunit of human liver y-glutamylcysteine synthetase and relative mRNA levels for heavy and light subunits in human normal tissues, Biochem. Biophys. Res. Comm. 206 (1995) 584-589. 36. R.T. Mulcahy and J.J. Gipp, Identification of a putative antioxidant response element in the 5'-flanking region of the human Y-glutamylcysteine synthetase heavy subunit gene, Biochem. Biophys. Res. Comm. 209 (1995) 227233. 37. R.T. Mulcahy, M.A. Wartman, H.H. Bailey and J.J. Gipp, Constitutive and b-naphthoflavone-induced expression of the human gamma-glutamylcysteine synthetase heavy subunit gene is regulated by a distal antioxidant response element/TRE sequence, J. Biol. Chem. 272 (1997) 7445-7454. 38. H.R. Moinova and R.T. Mulcahy, An electrophile responsive element (EpRE) regulates b-naphthoflavone induction of the human Y-glutamylcysteine synthetase regulatory subunit gene, J. Biol. Chem. 273 (1998) 14683-14689. 39. R. Gali, R and P.G. Board, Sequencing and expression of a cDNA for human glutathione synthetase, Biochemistry Journal 310 (1995) 353-358. 40. A.C. Wild and R.T. Mulcahy, Regulation of gamma-glutamylcysteine synthetase subunit gene expression: insights into transcriptional control of antioxidant defenses, Free Radic. Res. 32 (2000) 281-301. 41. E.L Dahl and R.T. Mulcahy, Cell-type specific differences in glutamate cysteine ligase transcriptional regulation demonstrate independent subunit control, Toxicol. Sci. 61 (2001) 265-272. 42. D.A. Dickinson, R.-M. Liu, K.E. lies and H.J. Forman, Signaling for the synthesis of glutathione. In: T. Yoshikawa, S. Toyokuni, Y. Yamamoto and Y. Naito (ed.), Free Radicals in Chemistry, Biology and Medicine. OICA International, London, 2000, pp. 139-154. 43. J.S. Woods and M.E. Ellis, Up-regulation of glutathione synthesis in rat kidney by methyl mercury. Relationship to mercury-induced oxidative stress, Biochem. Pharmacol. 50 (1995) 1719-1724. 44. Y. Urata, H. Yamamoto, S. Goto, H. Tsushima, S. Akazawa, S. Yamashita, S. Nagataki and T. Kondo, Long exposure to high glucose concentration impairs the responsive expression of Y-glutamylcysteine synthetase by interleukin-1b and tumor necrosis factor-alpha in mouse endothelial cells, J. Biol. Chem. 271 (1996) 1514615152.

12

H.J. Forman and D.A. Dickinson / Oxidative Signaling and Gluiathione Synthesis

45. T. Kondo, K. Yoshida, Y. Urata, S. Goto, S. Gasa and N. Taniguchi, y-Glutamyteysteine synthetase and active transport of glutathione S-conjugate are responsive to heat shock in K562 erythroid cells, J. Biol. Chem. 268 (1993)20366-20372. 46.1. Rahman, A. Bel, B. Mulier, M.F. Lawson, D.J. Harrison, W. MacNee and C.A.D. Smith, Transcriptional regulation of y-glutamylcysteine synthetase-heavy subunit by oxidants in human alveolar epithelial cells, Biochem. Biophys. Res. Comm. 229 (1996) 832-837. 47. D. Moellering, J. Me Andrew, R.P. Patel, H.J. Forman, R.T. Mulcahy, H. Jo and V.M. Dartey-Usmar, The induction of GSH synthesis by nanomolar concentrations of NO in endothelial cells: a role for gammaglutamylcysteine synthetase and gamma-glutamyl transpeptidase, FEBS Lett. 448 (1999) 292-296. 48. M.M. Shi, T. Iwamoto and H.J. Forman, y-Glutamylcysteine synthetase and GSH increase in quinone-induced oxidative stress in bovine pulmonary artery endothelial cells (BPAEC), Am. J. Physiol. 267 (1994) L414-L421. 49. R.M. Liu, H. Hu, T.W. Robison and H.J. Forman, Differential enhancement of y-glutamyl transpeptidase and yglutamylcysteine synthetase by tert-butylhydroquinone in rat lung epithelial L2 cells, American Journal of Respiratory Cell and Molecular Biology 14 (1996) 186-191. 50. D.C. Galloway and LI. McLellan, Inducible expression of the gamma-glutamytcysteine synthetase light subunit by t-butylhydroquinone in HepG2 cells is not dependent on an antioxidant-resoonsive element, Biochem. J. 336 (1998)535-539. 51. A.C. Wild and R.T. Mulcahy, Pyrrolidine dithiocarbamate up-regulates the expression of the genes encoding the catalytic and regulatory subunits of gamma-glutamytcysteine synthetase and increases intracellular glutathione levels, Biochem. J. 338 (1999) 659-665. 52. A.-L. Levonen, D.A. Dickinson, D.R. Moellering, R.T. Mulcahy, H.J. Forman and V.M. Dartey-Usmar, Biphasic effects of l5-deoxy-delta(12,l4)-prostaglandin J(2) on glutathione induction and apoptosis in human endothelial cells, Arterioscler. Thromb. Vase. Biol. 21 (2001) 1846-1851. 53. D.R. Moellering, A.-L. Levonen, Y.-M. Go, R.P. Patel, D.A. Dickinson, H.J. Forman and V.M. Dartey-Usmar, Induction of glutathione synthesis by oxidized low density lipoprotein and 1-palmitoyt-2-arachtdonylphosphatidyl choline: protection against quinone-mediated oxidative stress, Biochem. J. 362 (2002) 51-59. 54. R.-M. Liu, L. Gao, J. Choi and H.J. Forman, y-glutamylcysteine synthetase:

mRNA stabilization and

independent subunit transcription by 4-hydroxy-2-nonenal, Am. J. Physiol. 275 (1998) L861-L869. 55. R.-M. Liu, Z. Borok and H.J. Forman, 4-Hydroxy-2-nonenal increases gamma-glutamytcysteine synthetase gene expression in alveolar epithelial cells, Am. J. Respir. Cell. Mol. Biol. 24 (2001) 499-505. 56. K.S. Yao, A.K. Godwin, S.W. Johnson, R.F. Ozols, P.J. Odwyer and T.C. Hamilton, Evidence for altered regulation of gamma-glutamylcysteine synthetase gene expression among cisplatirvsensitive and cispiatinresistant human ovarian cancer cell lines., Cancer Res. 55 (1995) 4367-4374. 57. A. Gomi, T. Masuzawa, T. Ishikawa and M.T. Kuo, Posttranscriptional regulation of MRP/GS-X pump and yglutamylcysteine synthetase expression by 1-(4-amino-2-methyl-5-pyrimidinyl) methyl-3-(2-chloroethyl)-3nitrosourea and by cycloheximide in human glioma cells, Biochem. Biophys. Res. Comm. 239 (1997) 51-56. 58. R.F. Hamilton, Jr., M.E. Hazbun, C.A. Jumper, W.L. Eschenbacher and A. Holian, 4-Hydroxynonenal mimics ozone-induced modulation of macrophage function ex vivo, Am. J. Respir. Cell. Mol. Biol. 15 (1996) 275-282. 59. C.F. Babbs, Oxygen radicals in ulcerative colitis, Free Radic Biol Med 13 (1992) 169-181. 60. T.W. Robison, H.J. Forman and M.J. Thomas, Release of aldehydes from rat alveolar macrophages exposed in vitro to low concentrations of nitrogen dioxide, Biochim. Biophys .Acta 1256 (1995) 334-340. 61. M. Parola, G. Bellomo, G. Robino, G. Barrera and M.U. Dianzani, 4-Hydroxynonenal as a biological signal: Molecular basis and pathophysiological implications, Antioxidants and Redox Signaling 1 (1999) 255-284. 62. R.B. Tjalkens, S.W. Luckey, D.J. Kroll and D.R. Peterson, Alpha.beta-unsaturated aldehydes increase glutathione S-transferase mRNA and protein: correlation with activation of the antioxkJant response element, Arch. Biochem. Biophys. 359 (1998) 42-50. 63. J.Z. Cheng, R. Sharma, Y. Yang, S.S. Singhal, A. Sharma, M.K. Saini, S.V. Singh, P. Zimniak, S. Awasthi and Y.C. Awasthi, Accelerated metabolism and exclusion of 4-hydroxynonenal through induction of RLIP76 and

H.J. Forman and D.A. Dickinson / Oxidative Signaling and Glutathione Synthesis

hGST5.8 is an early adaptive response of cells to heat and oxidative stress, J. Biol. Chem. 276 (2001) 4121341223. 64. S. Spycher, S. Tabataba-Vakili, V.B. O'Donnell, L. Patomba and A. Azzi, 4-hydroxy-2,3-trans-nonenal induces transcription and expression of aldose reductase, Biochem. Biophys. Res. Comm. 226 (1996) 512-516. 65. S.E. Spycher, S. Tabataba-Vakili, V.B. O'Donnell, L. Palomba and A. Azzi, Aldose reductase induction: a novel response to oxidative stress of smooth muscle cells, FASEB J. 11 (1997) 181-188. 66. S.W. Luckey and D.R. Petersen, Metabolism of 4-hydroxynonenal by rat Kupffer cells, Arch. Biochem. Biophys. 389(2001)77-83. 67. S. Srivastava, S.Q. Liu, D.J. Conklin, A. Zacarias, S.K. Srivastava and A. Bnatnagar, Involvement of aldose reductase in the metabolism of atherogenic aldehydes, Chem. Biol. Interact. 130-132 (2001) 563-571. 68. S. Srivastava, A. Chandra, L.-F. Wang, W.E. Seifert, Jr, B.B. DaGue, N.H. Ansari, S.K. Srivastava and A. Bhatnagar, Metabolism of the lipid peroxidation product, 4-hydroxy-trans-2-nonenal, in isolated perfused rat heart, J. Biol. Chem. 273 (1998) 10893-10900. 69. M. Parola, G. Robino, F. Marra, M. Pinzani, G. Bellomo, G. Leonarduzzi, P. Chiarugi, S. Camandola, G. Poli, G. Waeg, P. Gentilini and M.U. Dianzani, HNE interacts directly with JNK isoforms in human hepatic stellate cells, J. Clin. Invest. 102 (1998) 1942-1950. 70. S. Camandola, A. Scavazza, G. Leonarduzzi, F. Biasi, E. Chiarpotto, A. Azzi and G. Poli, Biogenic 4-hydroxy-2nonenal activates transcription factor AP-1 but not NF-kappa B in cells of the macrophage lineage, Biofactors 6 (1997)173-179. 71. K. Uchida, M. Shiraishi, Y. Naito, Y. Torii, Y. Nakamura and T. Osawa, Activation of stress signaling pathways by the end product of lipid peroxidation. 4-hydroxy-2-nonenal is a potential inducer of intracellular peroxide production, J. Biol. Chem. 274 (1999) 2234-2242. 72. T. Herdegen and J.D. Leah, Inducible and constitutive transcription factors in the mammalian nervous system: control of gene expression by Jun, Fos and Krox, and CREB/ATF proteins, Brain Res. Rev. 28 (1998) 370-490. 73. C. Bonny, A. Oberson, S. Negri, C. Sauser and D.F. Schorderet, Cell-permeable peptide inhibitors of JNK: novel blockers of beta-cell death, Diabetes 50 (2001) 77-82. 74. B.L. Bennett, D.T. Sasaki, B.W. Murray, E.G. O'Leary, ST. Sakata, W. Xu, J.C. Leisten, A. Motiwala, S. Pierce, Y. Satoh, S.S. Bhagwat, A.M. Manning and D.W. Anderson, SP600125, an anthrapyrazolone inhibitor of Jun Nterminal kinase, Proc. Natl. Acad. Sci. U S A 98 (2001) 13681-13686. 75. L.M. Zipper and R.T. Mulcahy, Inhibition of ERK and p38 MAP kinases inhibits binding of Nrf2 and induction of GCS genes, Biochem. Biophys. Res. Comm. 278 (2000) 484-492.

14

Thiol Metabolism and Redox Regulation of Cellular Functions A. Pompella et al. (Eds.) IOS Press, 2002

Cell Survival and Changes in Gene Expression in Cells Unable to Synthesize Glutathione Emilio ROJAS*1, Zheng-Zheng SHI*, Mahara VALVERDE*, Richard S PAULES*, Geetha M. HABffi*, Michael W. LffiBERMAN* * Department of Pathology, Baylor College of Medicine, Houston, TX 77030 - email: [email protected];/NIEHS, cDNA Microrray Center, Research Triangle, Park, NC 27709; 1 Permanent address: Institute de Investigaciones Biomedicas, UNAM, Mexico D.F. 04510

1. A Genetic Approach to Glutathione Function Glutathione (GSH) is the major non-protein thiol in cells and can be present in cells and organs at concentrations as high as 8 mmol/g [1-3]. Most investigators have used standard biochemical and cell physiological approaches to investigate the function of glutathione About ten years ago we took a genetic approach and began to systematically clone enzymes involved in glutathione metabolism in the mouse. Subsequently, we developed mice deficient in many of these enzymes. This approach has great power because it allows one to uncover unanticipated functions. With respect to y-glutamyl cycle enzymes, it is worth noting that the approach has a minor drawback; mouse and human genes may differ in organization and even function. One example is mouse and human y-glutamyl transpeptidase (GGT); here the enzyme is encoded by a single gene in the mouse while in the human there are several genes and possible pseudogenes [4,5]. Nevertheless a clear picture of glutathione metabolism and the functions of individual enzymes has emerged from these genetic studies. It is known that glutathione is metabolized extracellularly, and in the mouse GGT is the only enzyme that catalyzes the removal of a y-glutamyl group [6]. GGT-deficient mice excrete large amounts of glutathione in their urine because in the absence of GGT the proximal tubules fail to cleave glutathione and thus to initiate the absorption of the constituent amino acids of this tripeptide. The result of the extensive loss of glutathione is the development of cysteine deficiency and stunted growth and failure of sexual maturation [6,7]. More recently we have identified and cloned another member of the GGT family termed y-glutamyl leukotrienase (GGL). This enzyme does not cleave glutathione and appears to have as its major substrate leukotriene €4, which is a glutathione derivative [8,9]. More recently we have developed mice deficient in GGL; these mice show no disturbances in glutathione metabolism, but are unable to convert leukotriene €4 to leukotriene D4 [10]. GGL appears to differ from a similar human enzyme termed GGT-rel (GGT-related) in that the latter enzyme is reported to metabolize glutathione on a limited basis [11]. We have performed similar studies on membrane-bound dipeptidase, an enzyme that is responsible in part for the cleavage of cysteinyl glycine [12,13].

E. Rojas et al. / Cell Survival and Changes in Gene Expression

15

2. Development of Mice Deficient in y-Glutamyl Cysteine Synthetase Glutathione is synthesized by the sequential action of y-glutamyl cysteine synthetase (y-GCS) and glutathione synthetase. In order to study the effects of lack of glutathione on growth and development, we developed mice deficient in the heavy subunit of y-glutamyl cysteine synthetase [10]. Mice homozygous for the deletion develop normally for the first few days of embryonic life but die sometime before E8.5 (See Figure 1). These mice fail to undergo gastrulation and mesoderm development and show substantial distal necrosis and interruption of DNA synthesis (Figure 2). These findings demonstrate that glutathione is essential for the growth and development of the murine embryo. Attempts to rescue these embryos by the administration of N-acetylcysteine to mothers were unsuccessful; however, y-GCS-deficient embryos maintained in culture continue to grow if exogenous glutathione or N-acetylcysteine (NAC) is added to the medium[10]. Although both in vivo and in embryo culture studies clearly demonstrate the importance of glutathione in growth and differentiation, they do not distinguish between functions unique to glutathione and those that can be supported by other thiols.

3. Development of Cells Deficient in y-GCS At the blastocyst stage, y-GCS-deficient embryos and wild-type embryos are morphologically indistinguishable. We reasoned that development is supported by GSH stored in the oocyte and that it might be possible to rescue y-GCS-deficient cells by the addition of glutathione to the culture medium. We have successfully isolated cell lines from homozygous mutant blastocysts by culturing them in medium containing GSH (Figure 3). These cell lines grow indefinitely in the presence of exogenous GSH and have cellular GSH levels as low as 2-10% of that in wild-type cells. We also found that N-acetylcysteine could effectively substitute for glutathione and that cells grew almost as well in NAC as in GSH, especially if y-GCS-deficient cell lines were reselected for growth in NAC (Figure 3). Such y-GCS-deficient cells grew for only a few days in the absence of GSH or NAC: cellular GSH drops to an undetectable level within 24 hours and cells eventually die. The importance of these findings is that they illustrate that the functions of GSH, but not GSH itself, are critical for cell growth in culture. This implies that GSH is not an essential co-factor for any enzyme necessary for continuous growth of these blastocyst-derived cells in culture. Further, differentiation and development are much more complicated processes than simple growth in culture. However, our in vivo experiments do not allow us to determine whether GSH itself or only its functions are required for development of a mammalian organism [10]. In order to evaluate the level of glutathione that will support cell growth, we measured glutathione concentrations in cells derived from wild-type blastocysts (BDC-1 cells) and yGCS-deficient cells (GCS-1 and GCS-1 nac) (Table 1). We found that GCS-1 cells growing in 2.5 mM glutathione contained less that 2% of the GSH found in wild-type cells (grown without the addition of GSH to the medium). Thus only a very small percentage of GSH normally found in cells is necessary for survival under culture conditions. Withdrawal of this GSH resulted in unmeasurable glutathione levels at 24 hours. When GCS-1 nac cells were maintained in medium lacking GSH but containing 2 mM NAC, the cells grew well (Figure 3) even though they had no detectable levels of GSH. Our results raise the question of why cells contain such a large excess of GSH and point to the importance of GSH in other cellular

16

E. Rojas el al. / Cell Survival and Changes in Gene Expression

Fig. 1. Developmental Abnormalities in E7.5 r-GCS-HS Mutant Embryos. (A and B) Whole-mount preparations of E7.5 normal and mutant embryos. The mutant embryos (B) are smaller than normal (A); note lack of organization in (B). (C-F) Histological comparison of normal (C and E) and mutant (D and F) embryos. The arrowheads in the sagittal sections (C and D) indicate the approximate position of the transverse sections (E and F). No mesoderm is apparent in the mutant embryos, (d) Deoidua; (ee) Embryonic ectoderm; (m) Mesoderm; (ve) Visceral endoderm. (G-H) Whole-mount in sfu hybridization analysis with a mesoderm marker, Brachyury (7). A normal expression pattern of T gene is shown in the wild type embryo (G). No signal was detected in a mutant littermate (H). Bar, 300 \un (A.B.G.H); 100 urn (C.D.E.F).

E. Rojas et al. / Cell Survival and Changes in Gene Expression

17

processes such as protection from oxidative stress and the detoxification and transport of toxic chemicals.

Fig. 2. In vivo Apoptosis and Proliferation in E6.5 y-GCS-HS Mutant Embryos. (A-B) Sagittal sections from 2 embryos [wild-type (WT) or heterozygous (+/-) and homozygous mutant (-/-)] were assayed by the TUNEL reaction. Fluorescein-labeled nuclei (orange) indicate apoptotic cells. Unlabeled nuclei appear red as a result of counterstaining with propidium iodide. The normal embryo (A) shows few apoptotic nuclei, whereas the mutant embryo (B) shows severe distal apoptosis (arrow). (C-D) Sagittal sections from WT or +/and -/- littermate embryos were analyzed by BrdU incorporation. Positive nuclei are visualized by green fluorescence. The mutant embryo (D) shows total absence of BrdU incorporation at its distal end (arrow), but the incorporation in other regions is comparable to the embryo (C). Genotypes of embryos were determined in adjacent sections by in situ hybridization with a y-GCS-HS exon 1 probe (not shown). Bar, 100 ^M.

4. Changes in Gene Expression after Glutathione Withdrawal Because GSH or its functions are essential for cell growth in culture, we reasoned that GSH withdrawal would produce changes in gene expression that might precede cell death. We have used three approaches in our initial studies. The first involves gene expression microarrays.

18

E. Rojas et al. / Cell Survival and Changes in Gene Expression

30-, 25-

U> mM GSH 2.0 mM NAC 1.25 mM MAC no addition

s is-i

1 2 3 4 5 Days after seeding Fig. 3. Growth of y-GCS-Deficient Cells. The open circles represent growth of a line of GCS-1 cells selected to grow in the presence of NAC (GCS-1 nac). Cells were seeded in M15 medium containing 15% of ES-cdl qualified fetal bovine serum and the additions shown. In all experiments, medium was changed daily.

Fig. 4. Comparison of RNA levels of genes that respond to the GSH withdrawal by Northern blotting. A) Comparison of Hexoquinase RNA levels between 24 and 48 hrs in the presence or absence of GSH. B) Comparison of Public Domain EST (GenBank # AA230989) RNA levels between 24 and 48 hrs in the presence or absence of GSH. Total RNA (10 tig/lane) was analyzed on the gel. These blots were stripped and reprobed with 18S cDNA to correct for loading differences.

E. Rajas et al. / Cell Survival and Changes in Gene Expression

Fig. 5. Myc family expression pattern in GCS-2 cells, in presence or absence of GSH for 24 hours. Gene expression was determined by ribonudease protection assay using 32P-labeled multiprobe template set. The antisense probe bands or probe line (PL) have a higher molecular weight than the hybridized probe fragments since the probe itself had not been protected during the RNase digestion.

20

E. Rojas et al. / Cell Survival and Changes in Gene Expression

We used an 8.8 K set of ESTs and cDNAs maintained by the NIEHS Microarray Center. In a second approach we have studied families of genes known to be involved in the cell cycle and cell death using ribonuclease protection. We have also begun an additional set of gene expression microarray experiments using a 12K set of Affymetrix mouse oligonucleotides We are confident that these approaches will ultimately lead to an extensive understanding of how GSH deficiency induces changes in gene expression in cultured cells. However, at present we have little concrete data. From the 8.8 K library, we have identified 32 candidate clones the expression of which either increases or decreases at least twofold 24 hours after the withdrawal of GSH. We have performed Northern blotting on about two-thirds of these clones, and in 87% of cases we have confirmed that the increases or decreases noted in the microarrays are real. Figure 4 illustrates the type of Northern data we have generated to show changes in gene expression. For ribonuclease protection analysis, we have made some educated guesses about what genes might logically be expected to increase or decrease as cells stop growing and die. An example of this approach is shown in Figure 5 in which we use a template set for genes related to cell cycle regulation. This rnMyc template set contains anti-sense RNA probes that can hybridize with target mouse mRNAs encoding sin3, C-myc, N-muy, L-myc, B-myc, max, mad, mxi, mad3, mad4, and mnt. Additional templates for the analysis of L32 and GAPDH housekeeping genes are included to allow assessment of total RNA levels for normalizing data. At present we have no firm conclusions about the spectrum of changes in gene expression that occurs when GSH falls to immeasurable levels in cells. There is some indication that there are changes in genes involving the cell cycle and apoptosis, but certainly a much broader range of changes is expected. Surveying these changes is a formidable task since a number of different time points need to be evaluated in order to understand the temporal sequence of changes. For the present our general strategy is to pick several time points and perform large-scale experiments. Following the identification of changes in the expression of individual genes, we will evaluate changes in their temporal expression using more traditional approaches such as northern blots and ribonuclease protection. One of the advantages of our approach is that it involves two complementary strategies. The screening approach that relies on microarrays makes no specific assumptions about which genes are involved and thus allows discovery. The use of ribonuclease protection assays involves some strategic guesses and is also useful since it is not now readily possible to screen the entire mouse genome by microarray technology. Another point of interest will be comparison of how changes in expression differ between y-GCS-deficient cells grown continuously in NAC (and therefore viable) and those deprived of GSH. Experiments of this type should allow us to distinguish between changes induced by impending cell death and those induced by the removal of GSH itself.

Acknowledgments Supported by NIH grants ES-07827 and ES-08668.

References 1. Fahey, R.C., Sundquist, A.R. Evolution of glutathione metabolism. In: Meister, A., ed. Advances in enzymology and related areas of molecular biology. New York: John Wiley & Sons, Inc., 1-53.1991.

E. Rojas et al. / Cell Survival and Changes in Gene Expression

2. Meister, A. Glutathione metabolism and its selective modification. J. Biol. Chem. 1998 263:17205-17208, 1988. 3. Meister, A., Anderson, M.E., Hwang, O. Intracellular cysteine and glutathione delivery systems. J. Am. Coll. Nutr. 5:137-151,1986. 4. Lieberman, M.W., Barrios, R., Carter, B. Z., Habib, G., Lebovitz, R.M., Rajagopalan, S., Sepulveda, A., Shi, Z.Z., and Wan, D.F.. Presidential Address: Y-Glutamyl transpeptidase: What does the organization and expression of a multipromoter gene tell us about its functions? Am.J.Pathol. 147:1175-1185,1995. 5. Chikhi, N., Holic, N., Guellaen, G. and Laperche.Y. y-Glutamyl transpeptidase gene organization and expression: a comparative analysis in rat, mouse, pig and human species. Comp. Biochem. Physiol.B Biochem. Mol. Biol. 122:367-80, 1999 6. Lieberman, M.W., Wiseman, A.L, Shi, Z.Z., Carter, B.Z., Barrios, R., Ou, C.N., Chevez-Barrios, P., Wang, Y., Habib, G.M., Goodman, J.C., Huang, S.L, Lebovitz, P.M., Matzuk, M.M. Growth retardation and cysteine deficiency in Y-glutamyl transpeptidase-deficient mice. Proc. Nat). Acad. Sci. USA. 93:7923-7926,1996. 7. Kumar, T.R., Wiseman, A.L., Kala, G., Kala, S.V., Matzuk, M.M., and Lieberman, M.W. Reproductive defects in y-glutamyl transpeptidase-deficient mice Endocrinology. 141:4270-4277, 2000. 8. Carter, B.Z., Shi, Z.Z., Barrios, R., and Lieberman, M.W. y-Glutamy! leukotrienase, a Y-glutamyl transpeptidase gene family member, is expressed primarily in spleen J. Biol. Chem. 273:28277-28285,1998. 9. Carter, B.Z., Wiseman, A.L, Orkiszewski, R., Ballard, K.D., Shields, J.E., Will, Y., Reed, D.J., Ou, C.N., and Lieberman, M.W. Metabolism of leukotriene 04 in Y-glutamyl transpeptidase-deficient mice. J. Biol. Chem. 272:12305-12310, 1997. 10. Shi, Z.Z., Osei-Frimpong, J., Kala, G., Kala, S.V., Barrios, R., Habib, G.M., Lukin, D.J., Danney, C.M., Matzuk, M.M., and Lieberman, M.W. Glutathione synthesis is essential for mouse development but not for cell growth in culture. Proc. Natl. Acad. Sci. USA 97:5101-5106, 2000. 11. Heisterkamp, N., Meyts, E.R-D., Uribe, L, Forman, H.J., and Groffen, J. Identification of a human Y-glutamyl cleaving enzyme related to, but distinct from, Y-glutamyl transpeptidase. Proc. Natl. Acad. Sci. USA 88:63036307, 1991. 12. Habib, G.M., Barrios, R., Shi, Z.Z., and Lieberman, M.W. Four distinct membrane-bound dipeptidase RNAs are differentially expressed in the mouse. J. Biol. Chem. 271:16273-16280,1996. 13. Habib, G.M., Shi, Z.Z., Cuevas, A., Guo, Q., Matzuk, M.M., and Lieberman, M.W. Leukotriene 04 and cystinylbis-glycine metabolism in membrane-bound-dipeptidase expression. Proc. Natl. Acad. Sci. USA 95:4859-4863, 1998.

22

Thiol Metabolism and Redox Regulation of Cellular Functions A. Pompella el al. (Eds.) IOS Press, 2002

Role of Glutathione in the Regulation of Liver Metabolism Jozsef MANDL and Gabor BANHEGYI Department of Medical Chemistry, Molecular Biology and Pathobiochemistry, Semmelweis University, Budapest, Hungary

1. Introduction Liver is a secretory organ, its basic function to export various molecules (plasma proteins, lipoproteins, glucose, ketone bodies, conjugated xeno- and endobiotics, glutathione, ascorbate, bile acids, cholesterol etc.) for utilisation by other organs or for the whole organism. The molecular weight of these compounds is different; large proteins and small molecules are equally involved. Luminal compartment of the endoplasmic reticulum is central site of these secretory processes. Previously it was supposed that mainly the secretory mechanism of macromolecules is tightly linked to the endomembrane system of the hepatocyte (for reviews see [1]). However, recently it has been suggested that the synthesis and export of smaller compounds is also related to the endoplasmic reticulum. Furthermore, luminal compartment of the ER is also important in final reactions proceeding the secretory processes. The special oxidative environment of the luminal compartment is essential to prove the necessary conditions for processes there. Oxidative environment is brought about by different transport mechanisms along the ER membrane. ER shares a central scene either in the intracellular metabolism and transport of the water-soluble antioxidants, mainly glutathione and ascorbate, or in the drug metabolism in hepatocytes. The oxidized/reduced glutathione ratio is markedly higher in the luminal compartment compared to the cytosol. Ascorbate/dehydroascorbate concentration is also higher in the luminal compartment, than in the cytosol. Moreover, their function in oxidative folding of proteins is fundamental. It is a question which molecules have regulatory functions in the complex control of processes in the luminal compartment of the endoplasmic reticulum. In the present paper attention is focused on connections between glycogen metabolism drug metabolism and intracellular redox state as they are related to each other in the luminal compartment of the ER in liver. Furthermore the regulatory functions of reduced glutathione is discussed.

2. Glycogen-Dependent Biosynthetic Pathways in the Liver. Possible Regulatory Role of Glutathione During the early phase of starvation the main carbohydrate reserve in the maintenance of blood glucose level is the hepatic glycogen, as glycogenolysis covers the glucose demand in the intermediary metabolism [2,3]. However, the function of hepatic glycogenolysis is not restricted only to the maintenance of the blood glucose level. The glycogenolysis

J. Mandl and G. Bdnhegyi / Regulation of Liver Metabolism

23

dependence of two other processes, drug glucuronidation [4,5] and ascorbate synthesis [6] has been described recently. The capacity of glucuronidation was decreased by inhibition of glycogenolysis caused by various agents (insulin, fructose, and glucose) in isolated hepatocytes from fed mice, which contain glycogen and also in hepatocytes from starved animals after glycogen depletion [5,7,8]. This phenomenon is a unique example of a special insulin glucagon synergism. In species, which did not lose their ability to synthesize ascorbate, similarly to glucuronidation, ascorbate production was also dependent on the extent of hepatic glycogen stores. Degradation of glycogen determined the rate of ascorbate formation: increase in glycogenolysis stimulated, while glycogenolysis inhibitors decreased the rate of ascorbate synthesis [6]. Thus, control of glycogen metabolism in the liver influences drug metabolism and intracellular redox homeostasis, too. Several observations suggested a special regulatory function of glutathione in control of glycogenolysis. Glutathione depletion dependent stimulation of glycogenolysis [9] caused an increase in glucuronidation [10] and ascorbate synthesis [11]. This is connected to the availability of UDP-glucuronic acid, the cofactor for glucuronidation and the precursor of ascorbate synthesis, which is predominantly originated from glycogen breakdown [12]. In accordance with this assumption, addition of UDP-glucose or UDP-glucuronic acid to permeabilised hepatocytes restored the capacity of both processes decreased by glycogen depletion/glycogenolysis inhibition [10,11,13]. It is concluded that glutathion through affecting glycogen metabolism has the ability to change the rate of glycogenolysis dependent drug glucuronidation and formation of ascorbate.

3. Glycogen Particles and the Luminal Compartment of the Endoplasmic Reticulum In hepatocytes sub-compartmentation of cytosolic glucose-6-phosphate pool has been supposed. Based on various observations it has been described that one of the glucose-6phosphate pools is functionally linked to glycogenolysis, while the other one is connected to gluconeogenesis [14,15]. The hypothesis on different glucose-6-phosphate pools and the supporting data are in agreement with the findings, that glucose uptake or gluconeogenesis are unable to cover the cofactor supply for glucuronidation and ascorbate production. The crucial question was arisen: what is the functional and morphological basis of this phenomenon (Fig.l.). It is tempting to conclude that the organisation of (smooth) endoplasmic reticulum bound multienzyme complexes responsible for the specific pathways/functions. UDPglucuronosyl transferases and gulonolactone oxidase similarly to glucose-6-phosphatase are integral membrane proteins/protein complexes of the endoplasmic reticulum [16,17]. The close association among glycogen particles and the vesicles and tubules of smooth endoplasmic reticulum has been already described in a study on the subcellular structure of hepatocytes [18]. Newly formed glycogen appears primarily in endoplasmic reticulum-rich regions of hepatocytes and remains associated with it during glycogen deposition and depletion [19] indicating that this subcellular structure is suitable for both glycogenesis or glycogenolysis, depending on the actual demands. A functional relationship is also indicated in von Gierke's disease in case of the glucose-6-phosphatase system. In addition to localisation the general state of glycogen metabolism is also correlated with the proliferation of the endoplasmic reticulum. During pregnancy large glycogen stores accumulate in the fetal liver, which are rapidly mobilised and depleted after the birth. In the

24

J. MandlandG. Bdnhegyi/Regulation of Liver Metabolism

GSH/GSSG

GSH glucose

G-6-P

UDP-glucuronate ox. drug

UDP-glucuronate.

drug glucuronide j drug glucuronide ascorbate

, -^^^

s

GLO]

dehydroascorbate

ER lumen cytosol dehydroascorbate Fig. 1. Connections between glycogen particles and iuminal compartment of the ER. Abbreviations (not used in the text): GSH: reduced glutathione, GSSG: oxidized glutathione, G-6-P-ase: glucose-6-phosphatase, G-6-P: glucose-6-phosphate, GLO: gulonolactone oxidase, ox: oxidised

fetal life the activity of glucose-6-phosphatase [20], UDP-glucuronosyltransferase [21] and gulonolactone oxidase (unpublished observation) is very low in the liver. Glycogenolysis and the glycogenolysis-dependent pathways seem to be activated together in the early postnatal period [22]. The proliferation of the smooth endoplasmic reticulum [23] and the induction of glucose-6-phosphatase system, several UDP-glucuronosyltransferase isoenzymes and gulonolactone oxidase occur simultaneously. A similar phenomenon appears during the fasting-refeeding cycle: at maximal glycogen accumulation the proportion of the smooth endoplasmic reticulum is very low in the hepatocyte, while intensive glycogenolysis is associated with the proliferation of the smooth fraction of endoplasmic reticulum [19,24].

4. ER Enzymes are supported by Transporters to make the Permeation of Substrates and Products Possible Latency is a characteristic feature of several endoplasmic reticulum enzymes in intact microsomal vesicles: their activity is low, while destruction of the membrane structure by detergents, pore-forming agents etc. results in an increases in their enzyme activity. The compartmentational or substrate-transport model of latency [3,16,25] suggests that the active site of these enzymes is intraluminal. Recent models for the membrane topology of glucose-6-phosphatase [26,27] and UDP-glucuronosyltransferases [16,21] verify the intraluminal positioning of the catalytic sites. Permeation of hydrophilic substrate(s) and

J. Mandl and G. Bdnhegyi/ Regulation of Liver Metabolism

25

product(s) to the lumen of the endoplasmic reticulum is mediated by specific transporters. Therefore, the velocity of the transport is rate limiting in the overall enzymatic process. The corresponding transport activities have been less explored, however, their existence has been proved. The transport of glucose-6-phosphate is evidenced by the demonstration of a metabolically active, intraluminal glucose-6-phosphate pool in microsomal vesicles [28]. The sequence of a glucose 6-phosphate translocase, mutated in glycogen storage disease type Ib, has also been published [29]. For the inward transport of UDP-glucuronic acid various antiport mechanisms have been suggested with the participation of UMP, UDP-Nacetylglucosamine or phenol glucuronides as counteranions [25,30,31]. The rapid permeation of gulonolactone through the microsomal membrane and the intraluminal accumulation of the products (ascorbate, hydrogen peroxide) of gulonolactone oxidase [32] suggests that this enzyme shares the orientation of glucose-6-phosphatase and UDPglucuronosyltransferases. Therefore, the final products of these glycogen-dependent pathways are formed in the lumen of endoplasmic reticulum, which is continuous in time with the extracellular environment. There are two possibilities for the export of these compounds: they can reach the plasma membrane by vesicular transport or can be secreted after two consecutive transport steps through the endoplasmic reticulum and plasma membranes. The molecules of smaller molecular mass seem to follow the second path, however, especially in the case of more charged and/or bulky compounds the participation of the vesicular transport cannot be excluded. Recent observations indicate that glucose [33] and ascorbate [34] exit from hepatocyte, at least partially, by this mechanism. The existence of endoplasmic reticulum transporters for the exit of products has been demonstrated. The permeation of the products of glucose-6-phosphatase, phosphate and glucose is mediated by T2 and T3, components of the glucose-6-phosphatase system [3]. Glucuronides and UMP, the end products of glucuronidation, exit by means of the above mentioned antiports. The product of gulonolactone oxidase can leave as ascorbate or as its oxidised derivative dehydroascorbate by newly described distinct transport mechanisms [33]. (The transport of dehydroascorbate is preferred, which is presumably mediated by the glucose transporter T3). Finally, the products intended for being used by other cells/organs should leave the hepatocyte by means of various transporters (multispecific organic anion transporter, GLUT2, organic anion transporting polypeptide 1, novel liver-secific transport protein etc.) on the canalicular or sinusoidal surface of the cell. What is the reason for the intraluminal organisation of these enzyme activities in the endoplasmic reticulum? A plausible explanation can be that the compartmentation of the hepatocellular substrate pool into a cytosolic and an intraluminal sub-pool (the latter is tightly connected to the pool in the glycogen particle by transporters) allows their independent regulation.

5. Conclusion: "The Glycogenoreticular System" and its Glutathione Dependent Regulation in the Liver Glucuronidation, ascorbate synthesis and glucose productions are dependent on hepatic carbohydrate reserves. These processes occur in the luminal compartment of the endoplasmic reticulum in hepatocytes. They have similar features: membrane bound enzymes with an intraluminal active site (glucose-6-phosphatase, UDPglucuronosyltransferases and gulonolactone oxidase) supported by transporters for the

26

7. Mandl and G. Bdnhegyi / Regulation of Liver Metabolism

membrane permeation of substrates (glucose-6-phosphate, UDP-glucuronate, gulonolactone) and products (glucose, glucuronides, ascorbate/dehydroascorbate). The final intended purpose of these liver-specific pathways is the export of the end products. In this context the glycogen particle can be regarded as the 'ribosome' of the smooth endoplasmic reticulum. On the basis of morphological and functional connections between hepatic glycogen and the (smooth) endoplasmic reticulum we propose to use the term "glycogenoreticular system" for the description of this export-orientated metabolic unit.

Mercapturic acid and cysteine conjugates ROS

liver toxicity GSH depletion

— GSH'

reactive intermediates NAPQI

acetaminophen

X

UDP glucuronate

\ UDP \glucuronate

AAP-glucuronide

ER lumen cytosol Fig.2. Acetaminophen induced liver injury and acetaminophen metabolism. Abbreviations: AAP: acetaminophen, ROS: reactive oxygen species, NAPQI: N-acetyl-pbenzoquinone-imine

The actual function of this "unit" is related to the current physiological/pathological state of the organism. The normal balance between these metabolic pathways is determined mainly by the complex regulation of glycogen metabolism, which ensures the priority of glycemic control and prevents those processes, which occur only under pathological circumstances and in diseases. Well known phenomena, as the depletion of glycogen stores caused by xenobiotics exposition or increased glycogenolysis due to the altered reduced/oxidised glutathione ratio changed by oxidant drugs, can be explained and interpreted this way. Thus, actual state of the antioxidant homeostasis determines the rate of glycogenolysis and indirectly regulates the actual function of the "unit". Defective expression of the components of the "unit" - subunits of the glucose-6-phosphatase system in the various subtypes of the von Gierke's disease, various forms of genetic polymorphism of bilirubin UDP-glucuronosyl transferase in Gilbert syndrome - alters the balance and coordination of glycogen metabolism and the connected pathways. The complex approach of the hepatic glycogenoreticular system may promote the better

J. Mandl and G. Bdnhegyi/Regulation of Liver Metabolism

27

understanding of pathophysiological states in which redox homeostasis, glucose metabolism and biotransformation are equally involved and the dominant role of glutathione. Paracetamol - a drug widely applied recently - toxicity gives a good example to demonstrate the function of the unit and its relationship to liver toxicity. Paracetamol can be conjugated in several ways mainly by glucuronidation (Fig.2.). However if glycogen is depleted and other metabolic forms - oxygenation - will be dominant [36], paracetamol will be more toxic, intracellular glutathione concentration as a known sensitive parameter of the oxidative damage by paracetamol derivatives, shows the actual state of toxicity [37].

Acknowledgments This work was supported by the Ministry of Health, by OTKA and by Hungarian Academy of Sciences (MTA).

References 1. Subcellular Biochemistry, Volume 21. Endoplasmic reticulum. M. Borgese and JR. Harris (eds.), Plenum Press, New York and London, 1993. 2. RC. Nordlie and RA. Jorgenson: Glucose-6-phosphatase. In: The Enzymes of Biological Membranes. A. Martonosi (ed.), Plenum Press, New York, 1976, Vol. 2., pp. 465-491. 3. YT. Chen and A. Burchell: Glycogen storage diseases. In: The metabolic basis of inherited disease. CR. Scriver, AL. Beaudet, WS. Sly, D. Valle (eds.) McGraw-Hill, New York, 1995, pp. 935-965. 4. RG. Thurman and FC. Kauffman: Factors regulating drug metabolism in intact hepatocytes. Pharmacoi. Rev. 31:229-251,1980. 5. G. Banhegyi, T. Garzo, F. Antoni, J. Mandl: Glycogenolysis - and not gluconeogenesis - is the source of UDP-glucuronic acid for glucuronidation. Btochim. Biophys. Acta 967:429-435, 1988. 6. L. Braun, T. Garzo, J. Mandl, G. Banhegyi: Ascorbic acid synthesis is stimulated by enhanced glycogenolysis in murine liver. FEBS Lett. 352:4-6,1994. 7. G. Banhegyi, R. Puskas, T. Garzb, F. Antoni, J. Mandl: High amounts of glucose and insulin inhibit pnitrophenol conjugation in mouse hepatocytes. Biochem. Pharmacoi. 42:1299-1302,1991. 8. J. Mandl, G. Banhegyi, MP. Kalapos, T. Garzo: Increased oxidation and decreased conjugation of drugs in the liver caused by starvation, (review) Chem.-Biol. Interact. 96:87-101,1995. 9. DM. Ziegler: Role of reversible oxidation-reduction of enzyme thiols-disulfides in metabolic regulation. Ann. Rev. Biochem. 54:305-329, 1985. 10. L Braun, G. Banhegyi, F. Puskas, M. Csala, T. Kardon, J. Mandl: Regulation of glucuronidation by glutathione redox state through the alteration of UDP-glucose supple originating from glycogen metabolism. Arch. Biochem. Biophys. 348:169-173, 1997. 11. L. Braun, M. Csala, A. Poussu, T. Garzo, J. Mandl: Glutathione depletion induces glycogenolysis dependent ascorbic acid synthesis in isolated murine hepatocytes. FEBS Lett. 388:173-176,1996. 12.G. Banhegyi, L. Braun, M. Csala, F. Puskas, J. Mandl: Ascorbate metabolism and its regulation in animals. Free Radio. Biol. Med. 23:793-803, 1997. 13.G. Banhegyi, T. Garz6, R. Fulceri, A. Benedetti, J. Mandl: Latency is the major determinant of UDPglucuronosyl-transferase activity in isolated hepatocytes. FEBS Lett. 328:149-152, 1993. 14. N. Kalant, M. Parniak, M. Lemieux: Compartmentation of glucose 6-phosphate in hepatocytes. Biochem. J. 248:927-931, 1987. 15. B. Christ, K. Jungermann: Sub-compartmentation of the 'cytosolic' glucose 6-phosphate pool in cultured rat hepatocytes. FEBS Lett. 221:375-380, 1987. 16.B. Burchell, MW. Coughtrie: UDP-glucuronosyltransferases. Pharmacoi. Ther. 43:261-289,1989. 17.K. Kiuchi, M. Nishikimi, K. Yagi: Purification and characterization of L-gulonolactone oxidase from chicken kidney microsomes. Biochemistry 21:5076-5082, 1982. 18. DW. Fawcett: Observations on the cytology and electron microscopy of hepatic cells. J. Natl. Cancer Inst. 15:1475-1503,1955. 19. RR. Cardell Jr: Smooth endoplasmic reticulum in rat hepatocytes during glycogen deposition and depletion. Int. Rev. Cytol, 48:221-279, 1977.

28

J. Mandl and G. Bdnhegyi / Regulation of Liver Metabolism

20. A. Leskes, P. Siekevitz, GE. Palade: Differentiation of endoplasmic reticulum in hepatocytes. I. Glucose6-phosphatase distribution in situ. J. Cell. Biol. 49264-287,1971. 21. DJ. Clarke, B. Burchell: Conjugation-Deconjugation Reactions in Drug Metabolism and Toxitity. (1994) In: Handbook of Experimental Pharmacology, Vol. 112 (Ed. Kauffman PC) Springer Veriag, Budapest, 1994, pp. 3-43. 22. J. Girard, JP. Pegorier: An overview of early post-partum nutrition and metabolism. Biochem. Soc. Trans. 26:69-74, 1998. 23. G. Dallner, P. Siekevitz, GE. Palade: Biogenesis of endoplasmic reticulum membranes. I. Structural and chemical differentiation in developing rat hepatocyte. J. Cell. Biol. 30:73-96,1966. 24. MB. Babcock, RR. Cardell Jr: Fine structure of hepatocytes from fasted and fed rats Am. J. Anat. 143:399438, 1975. 25. C. Berry, T. Hallinan: Summary of a novel, three-component regulatory model for uridine diphosphate glucuronyltransferase. Biochem. Soc. Trans. 4:650-652,1976. 26. W. Hemrika, R. Wever: A new model for the membrane topology of glucose-6-phosphatase: the enzyme involved in von Gierke disease. FEBS Lett. 409:317-319,1997. 27. CJ. Pan, KJ. Lei, B. Annabi, W. Hemrika, JY. Chou: Transmembrane topology of glucose-6-phosphatase. J. Biol. Chem. 273:6144-6148, 1998. 28. G. Bdnhegyi, P. Marcotongo, R. Fulceri, C. Hinds, A. Burchell, A. Benedetti: Demonstration of a metabolically active glucose-6-phosphate pool in the lumen of liver microsomal vesicles. J. Biol. Chem. 272:13584-13590, 1997. 29.1. Gerin, M. Veiga da Cunha, Y. Achouri, JF. Collet, E. Van Schaftingen: Sequence of a putative glucose 6phosphate transtocase, mutated in glycogen storage disease type Ib. FEBS Lett. 419235-238,1998. 30. X. Bossuyt, N. Blanckaert: Mechanism of stimulation of microsomal UDP-glucuronosyltransferase by UDPN-acetylglucosamine. Biochem. J. 305:321-328,1995. 31. G. Bdnhegyi, L Braun, M. Csala, P. Marcotongo, R. Fulceri, J. Mandl: Evidence for an UDP-glucuronic acid - phenol glucuronide antiport in rat liver microsomal vesicles. Biochem. J. 315:171-176,1996. 32. F. Puskas, L Braun, M. Csala, T. Kardon, P. Marcotongo, A. Benedetti, J. Mandl, G. Banhegyi: Gulonolactone oxidase activity-dependent intravesicular glutathtone oxidation in rat liver microsomes. FEBS Lett. 430:293-296,1998. 33. MT. Guillam, R. Burcelin, B. Thorens: Normal hepatic glucose production in the absence of GLUT2 reveals an alternative pathway for glucose release from hepatocytes. Proc. Natl. Acad. Sci. U S A . 95:1231712321.1998. 34. JM. Upston, A. Karjalainen, FL Bygrave: Efflux of hepatic ascorbate: a potential contributor to the maintenance of plasma vitamin C. Biochem. J. 342:49-56,1999. 35. G. Banhegyi, P. Marcotongo, F. Puskas, R. Fulceri, J. Mandl, A. Benedetti: Dehydroascorbate and ascorbate transport in rat liver microsomal vesicles. J. Biol. Chem. 2732758-2762,1998. 36. E. Evdokimova, H. Taper, P. Buc CakJeron: Role of ATP and glycogen reserves in both paracetamol sulfation and glucuronidatton by cultured precision-cut rat liver slices. Toxicol. In Vitro 15:683-690. 2001. 37. J. McClain, S. Price, S. Barve, R. Devalarja, S. Shedlofsky: Acetaminophen hepatotoxicity: an update. Curr. Gastroent. Rep. 1:42-49,1999.

Tlriol Metabolism and Redox Regulation of Cellular Functions A. Pompella el a I. (Eds.) [OS Press, 2002

29

Glutathione Transport in the Endo/Sarcoplasmic Reticulum Miklos CSALA*, Rosella FULCERI#, Jozsef MANDL*, Angelo BENEDETTI* andGaborBANHEGYI* ^Department of Medical Chemistry, Pathobiochemistry and Molecular Biology, Semmelweis University, H-1444, Budapest, FOB. 260, Hungary, and#Dipartimento di Fisiopatologia e Medicina Sperimentale, Universita di Siena, 53100 Siena, Italy

1. Introduction Glutathione has long been known as a major water-soluble antioxidant in animal tissues. It is now evident that glutathione (GSH) and its oxidised form, glutathione disulfide (GSSG) constitute the most important redox buffer both in the cytosol and in organelles. Each intracellular compartment can be characterized by a particular redox potential, which is reflected by the oxidation state of glutathione, that is the ratio of GSH and GSSG levels. In a typical mammalian cell, the ratio of [GSH]/[GSSG] in the cytosol is 30-100:1 resulting in a redox potential of about -230 mV. The lumen of the endoplasmic reticulum (ER) is more oxidized (-180 mV) with a 1-3:1 ratio of [GSH]/[GSSG] [1]. Although there is no direct experimental evidence, it is supposed that the lumen of the sarcoplasmic reticulum (SR) also has a higher redox potential and consequently a lower [GSH]/[GSSG] ratio than the cytosol. The maintenance of a higher redox potential of thiols and disulfides in the ER lumen is necessary for the oxidative protein folding in liver and other secretory organs [2-4] and is usually referred to as an oxidising environment. The proteins synthesised in the ER such as lysosomal, plasma membrane or secretory proteins are characterised by oxidised thiols, inta- and interchain disulfide bridges. The major function of SR is not the protein synthesis and processing but the storage and release of calcium. It seems that the glutathione redox buffer participates in the regulation of the latter too, as the redox potentials on the cytosolic and luminal surface of the SR membrane have a fundamental role in the regulation of calcium fluxes. A major mechanism for increasing cytosolic Ca2+ is the release of Ca2+ from internal stores via the members of a superfamily of intracellular calcium-release channels including ryanodine receptors (RyR) [5]. Hypersensitive thiols of RyRs are subjects of oxidoreduction, which cause the activation or inhibition of Ca2+ release [6-9]. Generally speaking, thiol oxidation by reactive oxygen species, glutathione disulfide (GSSG) and other thiol reagents activate, whilst reducing agents, such as glutathione (GSH), dithiothreitol and mercaptoethanol, inhibit the channel. Recent observations indicate a more sophisticated mechanism: RyR type 1 (RyRl) from skeletal muscle can function as a transmembrane redox sensor [10]. A large transmembrane redox potential inhibits, while dissipation of this potential activates the channel. Since glutathione is a hydrophilic charged compound that cannot diffuse through the biological membranes it seems likely that the transport of glutathione between the compartments plays a crucial role both in the maintenance and in the regulation of redox

30

M. Csala el al. / Glutathione Transport in the Endo/Sarcoplasmic Reticulum

conditions. Therefore, it is logical to consider that GSH/GSSG transport across the SR membrane is involved in regulating the local redox potential gradient necessary for the redox regulation of RyRl. In spite of its importance the transport of glutathione through endomembranes is a less explored field of cell biology. We have reported that GSH is transported through the membrane of hepatic ER at a relatively slow rate, while the membrane is practically impermeable towards GSSG [11]. Recent data showed that both compounds could permeate the membrane of SR vesicles from skeletal muscle, although with different velocity [10]. The protein(s) involved in glutathione transport remain to be identified. Here we report evidence for the involvement of RyRl in GSH/GSSG transport across the SR membrane of skeletal muscle. We observed that the initial rate of GSH and GSSG transport is higher in terminal cysternae vesicles, which have a higher relative abundance of RyRl. The activators and inhibitors of the RyRl calcium channel increase or decrease, respectively, the rate of glutathione transport. We suggest that RyRl may behave as a glutathione transporter on its own, or alternatively directly interact with a putative GSH/GSSG transporter.

2. Materials and Methods Preparation ofSR and ER Vesicles Microsomal vesicles were prepared from skeletal muscle, liver, brain and heart of New Zealand White rabbits (Milliner & Sons, Szilasliget, Hungary). SR membrane vesicles (total microsomal and purified terminal cisternae fractions) were prepared from the dominantly white hind limb skeletal muscles according to the method of Saito et al. [12]. Cardiac SR vesicles were isolated by the same procedure. Liver and brain ER vesicles were prepared as described earlier for rat liver [11]. Intactness of the vesicles was assessed by light scattering method (see below) using non-permeant compounds (i.e. sucrose, maltose, UDP-glucuronate). In the case of liver microsomes, membrane permeability was also confirmed by estimating the latency of the intravesicular enzyme UDPglucuronosyltransferase, which was higher than 95% [13]. The integrity of SR vesicles was also assessed on the basis of their ATP-dependent Ca * accumulation according to [14]. An almost complete calcium release was observed in terminal cisternae vesicles upon caffeine addition, indicating the purity of the fraction [14]. Microsomal preparations were frozen and maintained in liquid N2 until used. Transport Measurements by Rapid Filtration Method Rapid filtration experiments were executed as described in detail earlier [15,16]. Briefly, microsomal vesicles (1 mg protein /ml) were incubated in a buffer containing 100 mM KC1, 20 mM NaCl, 1 mM MgCl2, 20 mM MOPS, 1 mM GSH and its radiolabelled analogue [3H]GSH (10 u€i/ml) at 37°C. At the indicated times, vesicles were filtered through cellulose acetate-nitrate filter membranes (pore size 0.22 urn) and washed quickly on the filter with the same buffer containing 1 mM flufenamic acid, the inhibitor of GSH transport in ER [11] and SR [10] vesicles. The radioactivity retained on the filter was measured by liquid scintillation. Alamethicin (50 ng/mg protein) was included in parallel incubates to distinguish the intravesicular and the bound radioactivity. Alamethicin, a pore-forming antibiotic, makes the microsomal vesicles permeable towards various hydropbilic compounds such as UDP-glucuronate [13], sucrose, glucose-6-phosphate [15], GSH and GSSG [11]. The alamethicin-treated vesicles were recovered on filters and washed as

M. Csala et al. / Glutathione Transport in the Endo/Sarcoplasmic Reticulum

31

above. More than 95% of the microsomal .proteins was retained by the filters, indicating that alamethicin treatment did not affect the vesicular structure of microsomes as reported in [13]. The alamethicin-releasable portion of radioactivity (assumed as intravesicular) was calculated by subtraction. Transport Measurements by Light Scattenng Techniques Osmotically-induced changes in microsomal vesicle size and shape [17] were monitored at 400 nm at right angles to the incoming light beam, using a fluorimeter (Hitachi F-4500) equipped with a temperature-controlled cuvette holder (37°C) and magnetic stirrer. SR or ER vesicles (50 (ig/ml protein) were equilibrated for 2h in a hypotonic medium (5 mM KPIPES, pH 7.0). The osmotically-induced changes in light scattering were measured after the addition of a small volume (< 5%) of the total incubation volume of concentrated and neutralized solutions of the compounds to be tested as described in detail elsewhere [16]. Materials GSH, GSSG, ruthenium red, ryanodine, ATP, ADP, AMP, oleoyl-CoA were from Sigma, St. Louis, MO, USA. [3H]GSH was from NEN® Life Science Products, Inc., Boston, USA. All other chemicals were of analytical grade.

3. Results In the first set of experiments, the transport of radiolabelled GSH was measured in liver and skeletal muscle microsomes by a rapid filtration technique. The radioactivity associated with microsomes was measured in vesicles incubated both in the presence and absence of the pore-forming antibiotic, alamethicin to determine net intravesicular accumulation. The radioactivity associated to alamethicin-treated vesicles can be attributed to the binding of glutathione to the membrane or proteins. Alamethicin-permeabilized microsomes retained amounts of radioactivity less than 20% of that associated to untreated microsomes. Intravesicular GSH content was calculated as the alamethicin-releasable portion of the total radioactivity associated with the vesicles. The initial rates, time course and steady-state level of GSH uptake in liver microsomes were similar to our previous observations gained by an alternative method [11]. Both the extent and the rate of GSH uptake were significantly higher in muscle microsomes than in hepatic microsomes (Figure 1). The results gained by rapid filtration experiments might underestimate the rate of transport due to the unavoidable efflux of the investigated compound during the washing procedure. This discrepancy is obviously larger in case of a faster transport process. Therefore, the results were confirmed by permeability measurements using the light scattering method, which permits the real-time detection of the transport. The method is based on the detection of osmotic shrinkage and swelling of microsomal vesicles [17,1316]. Addition of non-permeable osmolytes causes a permanent shrinking of vesicles leading to a sustained increase in the light scattering signal. Permeant compounds cause a transient shrinking followed by a swelling phase as reflected by a gradual decrease in the light scattering signal. With highly permeant compounds, the transient shrinking phase may be small or even absent because of the very rapid equilibration of the compounds. In accordance with earlier observations [13-17], neither liver nor muscle microsomes nor isolated SR terminal cisternae vesicles were permeable to sucrose, maltose or UDPglucuronic acid while lower molecular weight compounds (glucose or KC1) rapidly entered the vesicles (data not shown). These observations as well, confirm the integrity of the

32

M. Csala etal. /Glutathione Transport in the Endo/Sarcoplasmic Reticulum

vesicle membranes. Addition of GSH or GSSG (6.25-25 mM) to liver microsomal vesicles leads to sustained increases in light scattering - the membrane is poorly permeable to them (Figure 2, traces L). Similar results were obtained with heart (Figure 2, traces H) and brain microsomes (Figure 2, traces B). In contrast, the addition of GSH or GSSG to SR vesicles hardly caused a shrinking phase indicating that both compounds crossed the membrane of SR vesicles rapidly (Figure 2, traces M), in agreement with the results of Feng et al. [10].

Rg. 1. Transport of GSH into liver and muscle microsomes detected by the rapid filtration method. Microsomal vesicles (1 mg protein/ml) were incubated in the presence of 3 mM GSH and tracer amounts of [3H]glutathk>ne (10 |iCi/ml) as described

in "Materials and Methods".

Microsomal vesicles permeabilized with alamethicin incubated

(50 in

ng/mg

parallel

protein)

were

experiments

to

evaluate radioactivity bound to microsomal membranes. At the indicated time points aliquots were withdrawn to measure 3H associated

with the

microsomes.

alamethicin-releasable

portions

The of

radioactivity (regarded as intravesicular) calculated by subtraction are shown. Data are means ± S.E.M. of 4-6 measurements. (A) liver microsomes; microsomes.

(•)

muscle

2

4

time (min)

Since the intra- and extravesicular concentrations of other components of the incubation medium had been equilibrated during a 2-h preincubation, the osmotically induced changes must be attributed to the movement of GSH or GSSG. Even higher permeability was observed in purified terminal cisternae - GSH and GSSG entered these vesicles instantly (Figure 2, traces TC). The permeability of the SR membranes was specific to GSH and GSSG: hydrophilic molecules of similar size to GSH/GSSG, such as sucrose, maltose and maltotetraose, did not enter muscle microsomal or terminal cisternae vesicles (data not shown). The permeability of the SR membrane towards GSH or GSSG was independent of the redox conditions: the addition of their various mixtures (GSH/GSSG from 50:1 to 1:1, 6.25 mM total concentration) resulted in similar light scattering traces (data not shown). Skeletal muscle microsomes and especially the subtraction enriched in terminal cisternae contain the RyR type 1. On the other hand, liver, heart and brain microsomes do not express this RyR isoform [5]. The above results indicate therefore a correlation between a high rate of GSH/GSSG transport and the presence of RyRl/In further experiments we studied the effect of RyR inhibitors [18,19] on SR permeability to GSH and GSSG. When muscle microsomes or terminal cisternae vesicles were incubated in the presence of 1-5 mM MgCli, glutathione (GSH or GSSG) influx was slower and its time course became similar to the transport observed in hepatic microsomes (Figure 3). The possible role of Cl"

M. Csala el at / Glutathione Transport in the Endo/Sarcoplasmic Reticulum

33

ions was ruled out by the addition of 2-10 mM KC1, which did not influence the permeability of the membranes to glutathione (data not shown). Similarly to Mg2+, addition of 2 JAM ruthenium red to the SR vesicles caused a dramatic inhibition of glutathione transport (Figure 3). Ryanodine was also inhibitory in micromolar concentrations; maximal

Fig. 2. Transport of GSH and GSSG Into liver, heart, brain and muscle mlcrosomes, and the terminal cistemae fraction detected by the light scattering method. Vesicles (50 ng protein/ml) were preequilibrated in a hyposmotic buffer. Osmotically-induced changes of light scattering following the addition of GSH (25 mM, panel a) or of GSSG (25 mM, panel b) were measured as described in 'Materials and Methods*. The osmolytes were added (arrow) to liver microsomes (L), heart microsomes (H), brain microsomes (B), to total muscle microsomal fraction (M) or to purified terminal cistemae fraction (TC). Representative traces are shown of 6-10 similar experiments on three different microsomal preparations.

O)

c 12 c 0) (/> CO

2> O c

O)

1 CO O CO

•*-•

O)

effect was reached at 200 uM (Figure 3). None of these agonists influenced the permeability of the membranes to glutamate, cysteine or to other small permeant compounds (i.e. glucose or phosphate) (data not shown). In a final set of experiments the effect of RyR channel activators on the glutathione transport in muscle microsomes was investigated. RyRl calcium channels can be activated by a variety of compounds including caffeine, adenine nucleotides and fatty acyl CoA esters (see 20 and 21 and refs therein). Moreover, fatty acyl CoA esters have been shown to counter the inhibitory effect of Mg2+ ions on RyR activity [21]. Under the present experimental conditions, transport of GSH or GSSG in SR vesicles is already so rapid that its increase would be hardly measurable. Therefore, muscle microsomes were preincubated in a buffer containing MgCh (1 mM) to maintain the transport of GSH or GSSG at slow rates (see Figure 3). Administration of adenine nucleotides (AMP, ADP or ATP) or oleoylCoA increased the influx rate of both compounds (Figure 4) but did not affect the permeability of the membranes to the other small compounds mentioned above (data not shown). Maximal stimulatory effect was observed at 2 mM in the case of adenine nucleotides and at 5 fiM in case of oleoyl-CoA (Figure 4). Addition of sucrose or maltose after these agents resulted in a sustained light scattering signal indicating that they did not

34

M. Csala et al. / Glutathione Transport in the Endo/Sarcoplasmic Reticulum

permeabilize the membrane (data not shown). The effect of caffeine cannot be clearly evaluated by the light scattering assay. Addition of caffeine (10 mM) caused, in fact, increases of both the signal and the background noise, possibly because the drug caused shrinking and/or aggregation of microsomal vesicles.

o> c

CD

none CO

0)

o> none Fig. 3. Effect of RyR antagonists on GSH and GSSG transport. The effect of Mg2* ions (Mg) ryanodine (Ry) and ruthenium red (RR) on GSH (panel a and c) or GSSG (panel b and d) transport was studied by the light scattering technique in total muscle microsomal fraction (panel a and b) and in a purified terminal cisternae fraction (panel c and d). MgCI2 (1 mM) was present in the hyposmotic buffer during the equilibration. Ryanodine (200 \M) and ruthenium red (2 \M) were added 2 minutes before GSH or GSSG. Representative traces are shown of 6-10 similar experiments on three different microsomal preparations.

4. Discussion As both reduced and oxidized forms of glutathione are hydrophilic and charged molecules, they require transporter proteins to cross the ER or SR membranes. The nature of these transporters is not yet clear. However, it is clear mat the transport activity for glutathione is different in liver ER as compared to muscle SR. In liver ER, only the reduced form of glutathione crosses the membrane [11]. In muscle, both GSH and GSSG can permeate the SR membrane [10, and the present study]. Moreover, the transport of GSH appears to be lower in liver ER than in muscle SR. The initial rate of radiolabelled GSH uptake was at least 4-fold lower in liver microsomes (Figure 1, earlier incubation time). In addition, the radiolabelled GSH taken up by liver microsomes is likely to be oxidized to GSSG and retained in this form inside the vesicles [11]. We have not yet evaluated the possible retention of GSSG by muscle microsomes, since it was not relevant in the present work. In muscle SR, however, the high permeability to GSSG suggests that it does not occur. Consistently, light-scattering measurements reveal that GSH transport is much more rapid in muscle total SR vesicles - and even more in the subfraction enriched in RyR channels -

M. Csala et al. / Glutathione Transport in the Endo/Sarcoplasmic Reticulum

35

than in liver microsomes. Moreover, light-scattering measurements reveal that little or no GSH/GSSG transport occurs in brain and heart microsomes; these are known to possess almost exclusively RyR channels other than the skeletal muscle isoform RyRl. It therefore appears that rapid transport of GSH/GSSG is restricted to ER/SR membranes enriched in RyRl.

Fig. 4.

Effect of RyR

agonists

on

GSH and

Total muscle microsomal vesicles were preincubated hyposmotic

containing

buffer

1 mM MgCI2.

GSH (panel a) or GSSG (panel b) transport was studied

by

the

light-

scattering technique. AMP, ADP or ATP (2 mM of each) or oleoyl-CoA (5 |xM) were added before GSH

2 minutes or GSSG.

Representative shown

traces are

of

6-10

similar

experiments

on

three

different preparations.

none

JZ

AMP

*h_

GSSG transport.

in

c J£ c

microsomal

0 CO CO

2 o c D)

0 & OS O V) •4—1

JC CD

ATP oleoylCoA

In skeletal muscle SR, it has been suggested that the GSH/GSSG permeability is in a functional relationship with the RyR channel in that it may contribute to the redox control of RyR channel-mediated Ca2+ fluxes [10]. Our present results indicate that the RyR activity can in rum directly control GSH and GSSG transport across the SR membrane. Testamental to this is that the transport of GSH and GSSG can be inhibited or activated by well-known inhibitors and activators of RyRl, respectively. This phenomenon appears to be specific for GSH and GSSG since the transport of other permeants in the SR membranes was unaffected by inhibitors/activators of RyRl. The function of the RyR channel in the generation of the Ca2+ signal during muscle contraction is well known. On the basis of our findings it is likely that passive GSH and GSSG fluxes are activated simultaneously with the release of intraluminal Ca2+ from the SR. The glutathione redox gradient between the cytosol and the lumen of the SR is allowed to equilibrate (at least partly). This phenomenon could play a role in altering the redox state of GSH and protein thiols of the skeletal muscle in contraction-induced injury [22]. Since the disappearance of the transmembrane redox gradient favors the open state of the calcium channel [10], the RyRl-dependent glutathione permeation may promote the mobilization of calcium in skeletal muscle.

36

M. Csala et al. / Glutathione Transport in the Endo/Sarcoplasmic Reticulum

As to the molecular nature of the skeletal muscle SR transporter of GSH and GSSG, it can be speculated that it is the RyRl itself. In particular, another channel can be formed by the supramolecular arrangement [23] of open RyRls in the presence of agonists allowing the permeation of glutathione. Alternatively, RyRl may be tightly coupled to the GSH/GSSG transport protein, and the activity of RyRl channel is directly transduced to the transporter resulting in a degree of co-regulation. This possibility is not unprecedented, since protein-protein interactions between the RyRl and the voltage-sensitive dihydropyridine receptor of the T tubule most likely transduce the activation of the latter to the former. Further work is needed to clarify these possibilities. Acknowledgements This work was supported by OTKA (National Scientific Research Fund) grants T32873 and F037484, a Hungarian Academy of Sciences Grant, a NATO linkage grant and a Telethon Grant No. 10602 to R. Fulceri. M. Csala was a recipient of a FEBS Short-Term Fellowship and a NATO Advanced Fellowship to Siena. Thanks are due to Dr. Roberta Giunti for the methodological advice and Mrs. Valeria Mile for her skilful technical assistance.

Abbreviations used RyR, ryanodine receptor; ER, endoplasmic reticulum; SR, sarcoplasmic reticulum; GSH, reduced glutathione; GSSG, glutathione disulfide; MOPS, 4-morpholinepropanesulfonic acid.

References [I] Hwang, C., Sinskey, A.J. and Lodish, H.F. (1992) Science 257,1496-1502 [2] Frand, A.R., Cuozzo, J.W. and Kaiser, C.A. (2000) Trends Cell Biol. 10, 203-210 [3] Tu, B.P, Ho-Schleyer, S.C., Travers, K.J. and Weissman, J.S. (2000) Science 290,1571-1574 [4] Cuozzo, J.W. and Kaiser, C.A. (1999) Nat. Cell Biol. 1,130-135 [5] Sorrentino, V., Barone, V. and Rossi, D. (2000) Curr. Opin. Gen. Oev. 10, 662-667 [6] Liu, G. and Pessah, I.N. (1994) J. Biol. Chem. 269, 33028-33034 [7] Zable, A.C., Favero, T.G. and Abramson, J.J. (1997) J. Biol. Chem. 272. 7069-7077 [8] Xia, R., Stangler, T. and Abramson, J.J. (2000) J. Biol. Chem. 275,36556-36561 [9] Sun, J., Xu, L, Eu, J.P., Stamter, J.S. and Meissner, G. (2001) J. Biol. Chem. 276,15625-15630 [10] Feng, W., Liu, G., Allen, P.O. and Pessah, I.N. (2000) J. Biol. Chem. 275, 35902-35907 [II] Banhegyi, G., Lusini, L, Puskas, F., Rossi, R., Fulceri. R., Braun, L. Mile, V., di Simplicio, P., Mandl, J. and Benedetti, A. (1999) J. Biol. Chem. 274,12213-12216 [12] Saito, A., Seiler, S., Chu, A. and Fleischer, S. (1984) J. Cell Biol. 99.875-885 [13] Fulceri, R., Banhegyi, G., Gamberucci, A., Giunti, R., Mandl, J. and Benedetti, A. (1994) Arch. Biochem. Biophys. 309, 43-46 [14] Fulceri, R., Giunti, R., Knudsen, J., Leuzzi, R., Kardon, T. and Benedetti. A. (1999) Biochem. Biophys. Res. Commun. 264, 409-412 [15] Banhegyi, G., Marcotongo, P., Fulceri, R., Hinds, C., Burchell. A. and Benedetti. A. (1997) J. Btol. Chem. 272, 13584-13590 [16] Banhegyi, G., Marcotongo, P., Puskas. F., Fulceri, R., Mandl, J. and Benedetti. A. (1998) J. Btol. Chem. 273, 2758-2762 [17] Meissner, G. (1988) Methods Enzymol. 157, 417-437 [18] Xu, L, Tripathy, A, Pasek, D.A. and Meissner, G. (1998) Ann. N. Y. Acad. Sci. 853,130-148

M. Csala et al. / Glutathione Transport in the Endo/Sarcoplasmic Reticulum

37

[19] Xu, L, Tripathy, A., Pasek, A.D. and Meissner, G. (1999) J. Biol. Chem. 274, 32680-32691 [20] Meissner, G. (1994) Annu. Rev. Physiol. 56, 485-508 [21] Fulceri, R., Knudsen, J., Giunti, R., Volpe, P., Mori, A. and Benedetti, A. (1997) Biochem. J. 325, 423-428 [22] McArdle, A., van der Meulen, J.H., Catapano, M., Symons, M.C., Faulkner, J.A. and Jackson, M.J. (1999) Free Radio. Biol. Med. 26, 1085-1091 [23] Yin, C.C. and Lai, F.A. (2000) Nat. Cell Biol. 2, 669-671

38

Thiol Metabolism and Redox Regulation of Cellular Functions A. Pompella et al. (Eds.) IOS Press, 2002

Role of Ascorbate in Oxidative Protein Folding Gabor BANHEGYI, Miklos CSALA, Angelo BENEDETTI*, Jozsef MANDL Dept. of Medical Chemistry, Mol. Biology and Pathobiochemistry, Semmelweis University, Budapest, Hungary; *Dip. di Fisiopatologia e Medicina Sperimentale, Univ. di Siena, Italy 1. Introduction According to theories on the origin of life, the first primitive living beings evolved in a reducing atmosphere. After the appearance of photosynthetic creatures, the atmosphere became oxidizing, but the presently existing organisms retained the memory of the ancient environment in the low redox potential of their cytoplasm. This yin-yang harmony between the oxidizing outer and reducing inner milieau is advantegeous for the organisms from the aspect of energy generation, but sometimes it may be dangerous because of the nonspecific, accidental oxidation reactions. Yin and yang principles mean the coexistence of the opposites; cells contain specific compartments for the purely oxidizing reactions to separate them from the cytoplasm, the main site of reductions. This way the formation of oxygen radicals is also localized. One of these compartmentalized reactions is the formation of intrachain and interchain disulfide bonds by the oxidation of protein-cysteine thiol groups (for recent reviews see [1-5]. The importance of disulfide bridge formation is evident because it plays a crucial role in the stabilization of the native conformation. Both in prokaryotic and eukaryotic cells, disulfide bond formation (oxidation and isomerization steps) are catalyzed exclusively in extracytoplasmic compartments. In eukaryotes, protein folding and disulfide bond formation are coupled processes that occur both co- and posttranslationally in the ER, which is the main site of the synthesis and posttranslational modification of secretory and membrane proteins. The formation of a disulfide bond from the thiol groups of two cysteine residues modifies the covalent structure of the polypeptide chain. The process requires the removal of two hydrogen atoms (two protons and two electrons), consequently, these bonds cannot form spontaneously; an oxidant is needed to accept the electrons. In aerobic conditions the ultimate electron acceptor is usually molecular oxygen; however, oxygen itself is not effective in protein thiol oxidation. Therefore, a more complex electron transfer chain from protein thiols to oxygen is supposed. In mammals the components of this putative electron transfer chain are mostly unknown. The present knowledge in this field comes from experiments on bacteria and yeast.

2. Oxidative Folding in Bacteria and Yeast It has been found that disulfide bond formation requires an enzyme called DsbA in the bacterial periplasm (an oxidizing compartment between the inner and outer bacterial membranes, functionally equivalent to the lumen of the ER) in Gram-negative bacteria (Fig. 1A). This protein accepts electrons from protein thiols and transfers them to an inner

G. Bdnhegyi el al./ Role of Ascorbute in Oxidative Protein Folding

39

membrane protein called DsbB. Oxygen was identified as the final electron acceptor in the process [6]. However, DsbB lacks cofactors, which can react with or can bind oxygen. Terminal oxidases of the bacterial electron transfer chain (cytochrome bd oxidase and cytochrome bo oxidase) are shown to act as last protein components of the chain and low molecular weight quinones (ubiquinone and menaquinone) also participate in the electron transfer [7]. On the other hand, a second pathway guarantees the reduction of misoxidized proteins. The periplasmic DsbC reduces proteins with incorrectly paired cysteine residues. Its active reduced state is maintained by the membrane protein DsbD. Electrons for the reduction of DsbD come from cytoplasmic NADPH with the mediation of thioredoxin (Fig. 1A). All Dsb proteins contain a thioredoxin-like fold and a CXXC motif. [8]

Fig A. Electron transfer connected to disulfide bond formation and to the reduction of Incorrect bonds In the perlplasm of E. coll (A) and In the ER of yeast (B). Dashed arrows indicate the direction of electron flow.

Similar pathways seem to promote the oxidative folding in yeast (Fig. IB). The thioredoxin-like protein protein disulflde isomerase (Pdilp) is the primary catalyst in disulfide bond formation [9]. Pdilp can be reoxidized by two alternative ways. A protein (Erolp) has been identified as an obligatory component of the protein thiol oxidizing machinery [10,11]. The enzyme is an intraluminal FAD-binding protein, which can oxidize Pdilp with oxidized FAD [12]. The ultimate electron acceptor is unknown, as the system is operative in both aerobic and anaerobic conditions. An additional mechanism is the Erv2p-mediated oxidation of

40

G. Bdnhegyi et ai/ Role ofAscorbate in Oxidative Protein Folding

Pdilp. Erv2p is a FAD-binding intraluminal protein, which uses molecular oxygen as a final electron acceptor [13-16]. However, only Erolp, but not Erv2p is essential for yeast growth. A flavin-containing monooxygenase (yFMO) can also play a role in the process. This enzyme uses molecular oxygen and NADPH to oxidize various small molecular weight thiols (among others GSH) on the cytoplasmic surface of the ER. The activity of the monooxygenase is induced severalfold in conditions that hinder thiol oxidation, i.e. in reductive stress [17,18]. However, it can be an additional rather than a basal mechanism, since the presence of the enzyme is not essential to cell viability. Moreover, GSSG transport has not been described in yeast ER, and GSH seems to compete with protein thiols in the oxidation process [19,20]. 3. Oxidative Folding in Mammalian Cells Although the detailed mechanism of oxidative folding has not been explored in mammalian cells, the oxidizing environment in the lumen of the ER is obviously required for the formation of disulfide bonds and for the proper folding of secretory proteins. The formation of native disulfide bonds is catalyzed by protein disulfide isomerase [21-24]. The mammalian analogues of Erolp keep PDI in active oxidized state [25-27]. The further stations on electron avenue are still unknown. The effect of unknown oxidant(s), necessary for the activity of enzymes participating in oxidative folding, is reflected in and supported by the glutathione redox buffer; the ratio of GSH and GSSG is around 2:1 within the lumen of ER and along the secretory pathway, whilst the cytosolic ratio ranges from 30:1 to 100:1 [28]. The continuously generated electrons can theoretically be eliminated by several possible mechanisms during protein thiol oxidation: 1.

Transmembrane electron transfer (e.g. b-type cytochromes in chromaffin granules and plant plasma membrane) [29,30].

2.

Preferential uptake of the oxidized member of a redox couple through the ER membrane and/or the efflux (or exocytosis) of its reduced form could also ensure the oxidative environment. For example, GSSG has long been thought to act as an electron acceptor, which reoxidizes PDI. In accordance with the hypothesis, preferential uptake of GSSG through the ER membrane was also suggested [28]. But in yeast cell line deficient in GSH synthesis disulfide bond formation was normal, and GSH can even compete with reduced proteins for oxidizing equivalents [19]. Moreover, GSSG transport was negligible in rat liver ER and the rate of GSH transport was much higher [31]. The preferential transport of dehydroascorbate (the oxidized form of ascorbate) described in rat liver microsomal vesicles [32] also supports the transport-based hypothesis.

3.

Enzymes (local oxidases) resident in the membrane or lumen of the ER can also play a role in disulfide bond formation. They can oxidize the thiol groups of proteins (sulfhydryl oxidases) directly. Alternatively, they can produce oxidizing compounds (e.g. reactive oxygen species; ROS) towards the lumen. Several microsomal enzymes (cytochrome P450s, NADPH cytochrome P450 reductase, gulonolactone oxidase [33,34], microsomal iron protein [35], NADPH-dependent oxidase, sulfhydryl oxidase [36-38], flavin-containing monooxygenase etc.) are able to produce ROS. ROS can oxidize protein thiols directly or by the mediation of electron carriers.

G. Bdnhegyi at at./ Role ofAscorbate in Oxidative Protein Folding

41

The above mechanisms are not exclusive, they can act even synergistically. Their common feature is the putative participation of ascorbate in the electron transfer. Therefore, we have hypothesised mat ascorbate (and presumably other low molecular weight antioxidants in the ER) participate in the electron transfer from protein thiols to oxygen.

4. Ascorbate and Electron Transfer The following observations and considerations suggest the participation of ascorbate (and/or other low molecular weight compounds) in the electron transfer chain: 1.

Analogous systems: quinones are components of mitochondrial and bacterial electron transfer chain in the terminal oxidation.

2.

Ascorbate (and quinones) can form an interface between oxidoreductions of different types mediating the transfer of one or two electrons.

3.

High concentration (and sometimes local synthesis) of ascorbate and other redox active compounds (quinones, tocopherol, glutathione) in the ER.

4.

Low representation of enzymatic antioxidants in the ER and in its neighborhood (e.g. the total absence of superoxide dismutase activity) despite the high representation of enzymes producing ROS.

5.

No significant protein thiol oxidation occurs in isolated microsomes; it suggests that a water-soluble diffusible factor from the cytosol is necessary for the process.

The role of ascorbate in the electron transfer from proteins has been suggested for a long time [39-41]. The known functions of ascorbate are based on its redox properties: it can be easily oxidized and reduced back; reactions with the transfer of one or two electrons are equally possible. Ascorbate can be connected to tocopherol and glutathione, other two abundant compounds of ER, by redox reactions. Topological coincidences also support the view that ascorbate can participate in electron transfer related to disulfide bond formation: the last steps of ascorbate synthesis are located in the ER [42]. Accordingly, ascorbate can be found in high concentration in the ER. Gulonolactone oxidase, an ER resident flavoenzyme catalyzing the final reaction of ascorbate biosynthesis, is known to produce the oxidant hydrogen peroxide as a byproduct [34]; ascorbate itself can also behave as a prooxidant under certain circumstance [43-45]. It has been observed that gulonolactone oxidase activity stimulated by gulonolactone addition results in the oxidation of GSH both in isolated hepatocytes and in microsomal systems [33]. The reaction led to intraluminal GSSG formation in GSH loaded microsomal vesicles [34]. On the basis of these observations we concluded that the dehydroascorbate/ascorbate redox couple must have an important role in the ER.

5. Dehydroascorbate-Dependent Thiol Oxidation in the Lumen of the ER Protein disulfide isomerase, a major protein in the ER lumen, is known to have a dehydroascorbate reductase activity by using GSH as an electron donor [46,47]. We have

42

G. Bdnhegyi el all Role ofAscorbate in Oxidative Protein Folding

observed that the rate of dehydroascorbate transport and intraluminal ascorbate accumulation was proportional with PDI activity in microsomes from various organs. High dehydroascorbate transport and intraluminal ascorbate accumulation was found in liver microsomes from BB/Wor spontaneously diabetic rats, which was due to the higher protein thiol levels [48]. The dehydroascorbate reductase activity of PDI was accompanied with the oxidation of protein thiols, indicating that the enzyme can use both GSH or protein thiols as the source of reducing equivalents. Accordingly, during protein thiol oxidation by dehydroascorbate, a simultaneous ascorbate formation was measured in rat liver microsomes.

6. Dehydroascorbate Transport in Rat Liver Microsomal Vesicles To act as an oxidant, dehydroascorbate (or its precursor ascorbate) must reach the luminal compartment of the ER. Therefore, membrane transporters for ascorbate and/or dehydroascorbate are also required. Ascorbate can derive from two sources in the lumen of the ER. In animals that express a functioning gulonolactone oxidase (e.g. rat) ascorbate is synthesized de novo in the lumen of the hepatic ER. Other cells are dependent on ascorbate uptake from the circulation. Xenogenous ascorbate (or its oxidized form, dehydroascorbate) has to be transported through the ER membrane from the cytosol in ascorbate-non-synthesizing species (e.g. guinea pig and human). Therefore, we have investigated ascorbate and dehydroascorbate transport by using two different experimental approaches: the rapid filtration and the light scattering techniques. It has been found that dehydroascorbate transport is favoured; its uptake quickly surpassed the level of the equilibrium. The transport was saturable, bidirectional, and temperature-dependent. It could be inhibited by high concentration of glucose and by typical glucose transport inhibitors (e.g. cytochalasin B, phloretin etc.). The results suggest that dehydroascorbate transport is mediated by a microsomal hexose transporter. On the other hand, ascorbate transport was negligible, did not reach the level of the passive equilibrium and it was saturated at very high ascorbate concentration [32]. The preferred transport of dehydroascorbate, its intraluminal reduction and the entrapment of ascorbate altogether can contribute to the generation of the luminal oxidizing environment.

7. Microsomal Ascorbate Oxidation The cytosolic concentration of dehydroascorbate is not known, but due to the reducing redoxpotential and the presence of dehydroascorbate reductases it is probably rather low. Therefore, a local microsomal ascorbate oxidation is supposed to supply the dehydroascorbate transporter with ligands. In fact, continuous ascorbate consumption could be observed in the presence of liver microsomes. Ascorbate oxidation resulted in a sustained level of ascorbyl free radical and dehydroascorbate till ascorbate was present in the medium. The highest level of ascorbate oxidase activity was found in liver microsomes. The enzyme catalyzing the reaction is presumably located on the outer surface of the microsomal vesicles, since protease treatment abolished its activity. The activity was not inhibited by the addition of enzymatic and nonenzymatic antioxidants, ruling out the participation of ROS in the mechanism. On the other hand, cytochrome and metalloprotein inhibitors effectively diminished ascorbate oxidation. The copper-specific neocuproine and other copper chelators were the most effective inhibitors of ascorbate oxidase activity, which may suggest that a copper-enzyme catalyses the reaction (unpublished observations).

G. Bdnhegvi et al./ Role of Ascorbate in Oxidative Protein Folding

43

8. Connections between Ascorbate Oxidation, Dehydroascorbate Transport and Protein Bisulfide Formation in Rat Liver Microsomes The tight coupling between ascorbate oxidation, dehydroascorbate transport and protein disulfide formation could be demonstrated by using the inhibitors of ascorbate oxidation (Fig 2A). The addition of these compounds hindered the transport of ascorbate (proving that ascorbate oxidation is a prerequisite for its microsomal transport; Fig. 2B) [49] and prevented the oxidation of intraluminal protein thiols (Fig. 2C) [50]. It should be noted that the mechanism of the ascorbate-dependent microsomal protein thiol oxidation is operative also in presence of the ascorbate-precursor gulonolactone in liver microsomes of species having gulonolactone oxidase activity and able to synthesize ascorbate. The gulonolactone-dependent microsomal protein thiol oxidation could be inhibited by the same compounds.

Fig. 2.

Connection between oxidative protein folding and ascorbate metabolism in the

endoplasmlc reticulum. Abbreviations: DMA, dehydroascorbate; ROS, reactive oxygen species; E vit, vitamin E; PDI, protein disulfide isomerase.

9. Possible Role of Other Antioxidants in the Electron Transfer Chain Liver microsomes are abundant in redox-active, lipophilic, low molecular weight compounds (e.g. tocopherol, ubiquinone, vitamin K etc.). Although they cannot promote protein thiol oxidation in the absence of an ultimate electron acceptor, they can participate in the electron transfer chain. Tocopheryl radical, for example, can be reduced by ascorbate that results in the regeneration of tocopherol, the active antioxidant. The possible role of tocopherol in the mechanism was investigated by measuring ascorbate-dependent protein thiol oxidation in liver microsomes from tocopherol-deficient rats. Although microsomal ascorbate oxidation was

44

G. Bdnhegyi et aU Role ofAscorbate in Oxidative Protein Folding

similar compared to the control, protein thiol oxidation was decreased in vitamin E-deficient microsomes. In vitro readdition of tocopherol partially restored protein thiol oxidation [51]. The results suggest that the absence of tocopherol uncouples ascorbate- and thiol oxidation. In case of tocopherol shortage the oxidizing power was used for lipid peroxidation. In summary, since ascorbate alone was able to promote protein thiol oxidation in rat liver microsomes, its dual role can be supposed. First, as an electron donor, it can activate oxygen (with the mediation of the ascorbate oxidase activity presumably due to a metalloprotein in the ER) and ROS can give rise to further dehydroascorbate generation. Second, the ascorbate / dehydroascorbate redox couple can transfer electrons between PDI and the other, presently unidentified components of the electron transfer chain. On the basis of our present knowledge and assumptions the following scenario can be outlined: an ascorbate oxidase activity attributable to a metalloprotein present in the microsomal vesicles oxidizes ascorbate to dehydroascorbate and generates ROS. ROS (directly or by the mediation of tocopherol) oxidize further ascorbate molecules. Dehydroascorbate (formed in or transported into the lumen of the ER) can be reduced by protein disulfide isomerase oxidizing the active center dithiols of the enzyme. Oxidized protein disulfide isomerase reacts with reduced nascent proteins yielding protein disulfides and catalytically regenerating protein disulfide isomerase. Although this scheme (Fig. 3) fits the results gained in in vitro microsomal systems, it is still questionable whether it describes correctly the in vivo situation. Further studies are needed to demonstrate whether the ascorbate/dehydroascorbate couple is the only mediator of the electron transfer or other compounds and mechanisms can replace it. Experiments in ascorbate-deficient cell cultures and in scorbutic animals will answer these important questions.

Fig. 3. Effect of various compounds on ascorbate oxidation (A), ascorbate transport (B) and intraluminal protein thiol oxidation (C) In rat liver microsomes. Abbreviations: eco, econazote; pro, proadifen; q, quercetin; neoc, neocuproine. Inhibitors were used at 100 //M concentration.

G. Bdnhegyi et al./ Role ofAscorbate in Oxidative Protein Folding

45

Acknowledgements Experimental work summarized in this paper was supported by the Hungarian Scientific Research Fund (T32873), by the Hungarian Academy of Sciences and by the Hungarian Ministry of Health (ETT 242/2000). Italian-Hungarian cooperation was supported by grants from NATO, CNR and the Hungarian Science and Technology Foundation. References 1. Debarbieux, L, and Beckwith, J. (1999) Electron avenue: pathways of disulfide bond formation and isomerization. Cell 99,117-119 2. Frand, A.R., Cuozzo, J.W., and Kaiser, C.A. (2000) Pathways for protein disulphide bond formation. Trends Cell Biol. 10, 203-210 3. Freedman R.B., Dunn, A.D., and Ruddock, LW. (1998) Protein folding: A missing redox link in the endoplasmic reticulum. Curr. Biol. 8, R468-R470 4. Glockshuber, R. (1999) Where do the electrons go? Nature 401, 30-31 5. Kadokura, H., and Beckwith, J. (2001) The expanding world of oxidative protein folding. Nat. Cell Biol. 3, E247E249 6. Bader, M., Muse, W., Zander, T., and Bardwell, J.C.A. (1998) Reconstitution of a protein disulfide catalytic system. J. Biol. Chem. 273, 10302-10307 7. Bader, M., Muse, W., Ballou, D.P., Gassner, C., and Bardwell, J.C.A. (1999) Oxidative protein folding is driven by the electron transport system. Cell 98, 217-227 8. Fabianek, R.A., Hennecke, H., and Thony-Meyer, L. (2000) Periplasmic protein thiol:disulfide oxidoreductases of Escherichia coli. FEMS Microbiol. Rev. 34, 303-316 9. Frand, A.R., and Kaiser, C.A. (1999) Erolp oxidizes protein disulfide isomerase in a pathway for disulfide bond formation in the endoplasmic reticulum. Molec. Cell 4, 469-477 10. Frand, A.R., and Kaiser, C.A. (1998) The ERO1 gene of yeast is required for oxidation of protein dithiols in the endoplasmic reticulum. Molec. Cell 1,161-170 11. Pollard, M.G., Travers, K.J., and Weissman, J.S. (1998) Erolp: a novel and ubiquitous protein with an essential role in oxidative protein folding in the endoplasmic reticulum. Molec. Cell 1,171-182 12. Tu, B.P., Ho-Schleyer, S.C., Travers, K.J., and Weissman, J.S. (2000) Biochemical basis of oxidative protein folding in the endoplasmic reticulum. Science 290,1571-1574 13. Gerber, J., Miihlenhoff, U., Hofhaus, G., Lilt, R., and Lisowsky, T. (2001) Yeast Erv2p is the first microsomal FAD-linked sulfhydryl oxidase of the Erv1p/Alrp protein family. J. Biol. Chem. 276, 23486-23491 14. Lee, J.-E., Hofhaus, G., and Lisowsky, T. (2000) Ervlp from Saccharomyces cerevisiae is a FAD-tinked sulfhydryl oxidase. FEBS Lett. 477, 62-66 15. Sevier, C.S., Cuozzo, J.W., Vala, A., Aslund, F., and Kaiser, C.A. (2001) A flavoprotein oxidase defines a new endoplasmic reticulum pathway for biosynthetic disulphide bond formation. Nat. Cell Biol. 3, 874-882 16. Gross, E., Sevier, C.S., Vala, A., Kaiser, C.A., and Fass, D. (2001) A new FAD-binding fold and irrtersubunit disulfide shuttle in the thiol oxidase Erv2p. Nat. Struct. Biol. 9, 61-67 17. Suh, J.-K., Poulsen, L.L., Ziegler, D.M., and Robertas, J.D. (1999) Yeast flavin-containing monooxygenase generates oxidizing equivalents that control protein folding in the endoplasmic reticulum. Proc. Natl. Acad. Sci. USA 96, 2687-2691 18. Suh, J.-K., and Robertus, J.D. (2000) Yeast flavin-containing monooxygenase is induced by the unfolded protein response. Proc. Natl. Acad. Sci. USA 97, 121-126 19. Cuozzo, J.W., and Kaiser, C.A. (1999) Competition between glutathione and protein thiols for disulphide-bond formation. Nat. Cell Biol. 3,130-135

46

G. Bdnhegyi et aU Role ofAscorbate in Oxidative Protein Folding

20. Bader, M., Winther, J.R., and Bardwell, J.C.A. (1999) Protein oxidation: prime suspect found 'not guilty". Nat CellBiol. 1.E56-E58 21. Gilbert, H.F. (1997) Protein disulfide isomerase and assisted protein folding. J. Bid. Chem. 272, 29399-29402 22. Ferrari, D.M., and Soling, H.-D. (1999) The protein disulphide-isomerase family: unravelling a string of folds. Biochem. J. 339,1-10 23. Molinari, M., and Hetenius, A. (1999) Glycoproteins form mixed disulphides with oxidoreductases during folding in living cells. Nature 402, 90-93 24. Bulleid, N.J., and Freedman, R.B. (1988) Defective co-translational formation of disulphide bonds in protein disulphide-isomerase-deficient microsomes. Nature 335, 649-651 25.Cabibbo, A., Pagani, M., Fabbri, M.. Rocchi, M., Farmery, M.R., Bulleid, N.J., and Sitia, R. (2000) ERO1-L, a human protein that favors disulfide bond formation in the endoplasmic reticulum. J. Biol. Chem. 275,4827-4833 26. Pagani, M., Fabbri, M., Benedetti, C., Fassio, A., Pilati, S, Bulleid, N.J., Cabibbo. A., and Sitia, R. (2000) Endoplasmic reticulum oxidoreductin 1-la (ERO1-La), a human gene induced in the course of the unfolded protein response. J. Biol. Chem. 275, 23685-23692 27. Benham, A.M., Cabibbo, A., Fassio. A.. Bulleid, N., Sitia, R.. and Braakman, I. (2000) The CXXCXXC motif determines the folding, structure and stability of human Erol-Lct. EMBO J. 19, 4493-4502 28. Hwang, C., Sinskey, A.J., and Lodish, H.F. (1992) Oxidized redox state of glutathione in the endoplasmic reticulum. Science 257,1496-1502 29. Asard, H., Horemans, N., and Caubergs, R.J. (1992) Transmembrane electron transport in ascorbate-loaded plasma membrane vesicles involves a b-type cytochrome. FEBS Lett. 306,143-146 30. Horemans, N., Foyer, C.H., and Asard, H. (2000) Transport and action of ascorbate at the plant plasma membrane. Trends Plant Sci. 5, 263-267 31. Banhegyi, G., Lusini, L, Puskas, F., Rossi, R., Fulceri, R., Braun, L, Mile, V., di SimpUck), P., Mandl, J., and Benedetti, A. (1999) Preferential transport of glutathione versus glutathione disulfide in rat Hver microsomal vesicles. J. Biol. Chem. 274,12213-12216 32. Banhegyi, G., Marcotongo, P., Puskas, F., Fulceri. R., Mandl, J., and Benedetti. A. (1998) Dehydroascorbate and ascorbate transport in rat liver microsomal vesicles. J. Biol. Chem. 273, 2758-2762 33. Banhegyi, G., Csala, M., Braun, L, Garzd, T., and Mandl, J. (1996) Ascorbate synthesis-dependent glutathione consumption in mouse liver. FEBS Lett. 381, 39-41 34. Puskas, F., Braun, L, Csala, M., Kardon, T., Marcotongo, P., Benedetti, A.. Mandl, J., and Banhegyi, G. (1998) Gutonolactone oxidase activity-dependent intravesicular glutathione oxidation in rat liver microsomes. FEBS Lett. 430, 293-296 35. Minotti, G., and Ikeda-Saito, M. (1991) Bovine heart microsomes contain a Mr = 66,000 non-heme iron protein which stimulates NADPH oxidation. J. Biol. Chem. 266, 20011-20017 36. Hoober, K.L, and Thorpe, C. (1999) Egg white sulfhydryl oxidase: kinetic mechanism of the catalysis of disulfide bond formation. Biochemistry 38, 3211-3217 37. Hoober, K.L, Sheasley, S.L, Gilbert, H.F., and Thorpe, C. (1999) Sulfhydryl oxidase from egg white. A facile catalyst for disulfide bond formation in proteins and peptides. J. Biol. Chem. 274, 22147-22150 38. Janolino, V.G., and Swaisgood, H.E. (1987) Sulfhydryl oxidase-catalyzed formation of disulfide bonds in reduced ribonuclease. Arch. Biochem. Biophys. 258, 265-271 39. Szent-Gyorgyi, A. (1978) The Living State and Cancer (Marcel Dekker. New York-Basel) 40. Venetianer, P., and Straub, F.B. (1964) The mechanism of action of the ribonudease-reactrvating enzyme. Biochim. Biophys. Acta89, 189-190 41. Venetianer, P., and Straub, F.B. (1965) Studies on the mechanism of action of the ribonudease-reactivating enzyme. Acta Physiol. Acad. Sci. Hung. 27, 303-315 42. Banhegyi, G., Braun, L, Csala, M., Puskas, F., and Mandl, J. (1997) Ascorbate metabolism and its regulation in animals (review). Free Radic. Biol. Med. 23, 793-803 43. Carr, A., and Frei, B. (1999) Does vitamin C act as a pro-oxklant under physiological conditions? FASEB J. 13, 1007-1024

G. Bdnhegyi et al./ Role ofAscorbate in Oxidative Protein Folding

47

44. Halliwell, B. (1999) Vitamin C: poison, prophylactic or panacea? Trends Biochem Sci. 24, 255-259 45. Smirnoff, N. (2000) Ascorbic acid: metabolism and functions of a multi-fecetted molecule. Curr. Opin. Plant Biol. 3, 229-235 46. Wells, W.W., Xu, D.P., Yang, Y., and Rocque, P.A. (1990) Mammalian thioltransferase (glutaredoxin) and protein disulfide isomerase have dehydroascorbate reductase activity. J. Biol. Chem. 265,15361-15364 47. Wells, W.W., and Xu, D.P. (1994) Dehydroascorbate reduction. J. Bioenerg. Biomembr. 26, 369-377 48. Nardai, G., Braun, L, Csala, M., Mile, V., Csermely, P., Benedetti, A., Mandl, J., and Banhegyi, G. (2001) Protein disulfide isomerase and protein thiol dependent dehydroascorbate reduction and ascorbate accumulation in the lumen of the endoplasmic reticulum. J. Biol. Chem. 276, 8825-8828 49. Csala, M., Mile, V., Benedetti, A., Mandl, J., and Banhegyi, G. (2000) Ascorbate oxidation is a prerequisite for its transport into rat liver microsomal vesicles. Biochem. J. 349, 413-415 50. Csala, M., Braun, L., Mile, V., Kardon, T., Szarka, A., Kupcsulik, P., Mandl, J., and Banhegyi, G. (1999) Ascorbate mediated electron transfer in protein thiol oxidation in the endoplasmic reticulum. FEBS Lett. 460, 539-543 51. Csala, M., Szarka, A., Marginal, §., Mile, V., Kardon, T, Braun, L, Mandl, J., and Banhegyi, G. (2001) Role of vitamin E in ascorbate-dependent protein thiol oxidation in rat liver endoplasmic reticulum. Arch. Biochem. Biophys. 388, 55-59

48

Thiol Metabolism and Redox Regulation of Cellular Functions A. Pompella et al. (Eds.) IOS Press, 2002

Cytophotometric Investigations on Oscillating Thiol-Disulfide Equilibria and Oxidized Protein Sulfur Gerhard NOHAMMER Institute of Molecular Biology, Biochemistry and Microbiology (IMBM), Karl-FranzensUniversity ofGraz, A-8010 Graz, Heinrichstrasse 31A, Austria - Fax: 00433163809016; Email: gerhard. noehammer@kfunigra-. ac. at 1. Introduction Cellular thiol-disulfide equilibria reflect the response of cells to distinct metabolic situations. The state of oxidation of protein thiols of most cysteine- (and methionine)containing proteins influences their structure and function. Cellular redox-systerns exert important effects on the redox-state of cellular thiols. Among these, the gluthathionegluthathione-disulfide-system is one of the most important, connected with gluthathione consuming and regenerating systems and involved consequently in activation and deactivation of many enzymes by posttranslational modification [1], in protein synthesis, and in receptor binding [2]. Oxidative stress might be seen as a situation challenging the cellular redox-systems to meet it by using also reactive oxygen species for e.g. activation of stress-inducible genes [3]. Tumor-associated changes of protein thiols and disulfides in normal tissue, providing evidence for the existence of a biochemical «field effect» and an «extented field effect» of malignant tumors [4,5,6,7] require methods more convenient for the histochemical demonstration of protein thiols [8] and disulfides [9] than the DDD-(2,2'dihydroxy-6,6'-dinaphthyldisulfide)-Fast blue B-methods used [10]. Metal salts (e.g. Ag+, Hg2+) react with disulfides forming 1.5 moles of mercaptide and sulfinic acid per mole of disulfide [11]. The aim of the present work was to use this reaction of mercuric salt in combination with the disulfide reagent DDD for the histochemical demonstration of protein disulfides. 2. Materials and Methods Histochemical Methods The cells used, e.g. Ehrlich ascites tumor cells (EATC) and Yoshida ascites tumor cells (YOATC), were washed twice with 0.15M NaCl (4°C) by centrifugation (1500 rpm; 5 min) and the washed cell pellet, resuspended in 0.15M NaCl, used for fixation, performed using different methods: 1) One drop each of the concentrated cell suspension was smeared on a glass-slide, indexed by ascending numbers. The smeared cells were fixed immediately by a 2 sec spray-fixation using ether/ethanol (1:1) followed by postfixation in methanol (4°C). This procedure requires approximately 15 sec per slide and can be performed with more than one slide at a time.

G. Nohammer/ Cytophotometric Investigations

49

2) Aliquots of the washed cell suspension (or all of) were pipetted into excess methanol (4°C) at distinct times. After at least 30 min fixation in methanol, the fixed cell suspension was centrifuged (1500 rpm; 5 min), the pellet resuspended in 0.15M NaCl, centrifuged again and finally the cell pellet was resuspended in 0.15M NaCl to obtain a concentrated cell suspension used for cell preparations either as descibed above ,or using the cytospin (Shandon Southern; 900 rpm; 7 min). Generally, the cell preparations were postfixed and stored in cold methanol. DDD-Fast Blue B-Staining Fixed cell preparations, sometimes pretreated as decribed later on, were put into a DDDsolution (100 mg 2,2'-dihydroxy-6,6'-dinaphthyldisulfide [DDD] are dissolved in 55 ml ethanol (95%) and diluted then with different buffers to enable DDD-reactions to run at different pHs {The buffers used were, 0.1 M acetic acid adjusted to pH 2.5, 3, 4, and 5 respectively; 0.1M Tris pH 6 and 7; barbital-acetate buffer pH8.6 [12 ].} ). To stop the DDD-reaction at distinct times and to remove excess DDD, the cell preparations were put into acetone for 5 min followed by 3x10 min acetone. The washed cell preparations were transferred into distilled water for 5 min, and the protein bound 2-hydroxy-6-thionaphthol was then stained using a freshly prepared solution of 100 mg Fast blue B (FB) in 100 nil of 0.1M phosphate buffer pH 6.5 for a 10 min coupling-reaction. Excess FB was removed by 5 min running tap water. The stained cell preparations were dehydrated by an ascending series of ethanol, transferred to xylene and, still wet with xylene, embedded using Merckoglas. Pretreatments of Fixed Cell Preparations 1) Preincubation with mercuric acetate (Hg//): Fixed cell preparations were put into a freshly prepared solution of 319 mg mercuric acetate (HgAc) in 50 ml 0.1 M acetate buffer pH4 and diluted then with 50 ml ethanol (95%). Incubation was stopped by washing the preparations 2x5 min in a mixture of 0.1M acetate buffer pH4 and ethanol (1:1) followed by 2x5 min 50% ethanol. When cell preparations were preincubated with HgAc followed by a DDD-reaction the histochemical method has been designated Hg//DDD. 2) Preincubation with 2,4-dinitrofluorobenzene (2,4/7): To block protein thiols (and other nucleophilic residues) irreversibly fixed cell preparations were placed for one day into a freshly prepared solution of 2,4-dinitrofluorobenzene (2,4-DNFB)( 175 ^1 2,4-DNFB dissolved in 50 ml ethanol (95%) and diluted then with 50 ml 0.1 M Tris pH 7.4). Excess 2,4-DNFB was removed by 4x5 min washing in 50% ethanol. When cell preparations were preincubated with 2,4-DNFB followed by the DDD-reaction the histochemical method has been designated 2,4//DDD. When 2,4/7 was followed by pretreatment with HgAc prior to the DDD-reaction, the histochemical method has been designated 2,4//Hg//DDD. 3) Preincubation with cysteine: To reduce oxidized protein sulfur species under mild conditions, fixed cell preparations were put into a 0.2M solution of cysteine in 0.1 M acetate buffer pH4. Reduction was stopped and excess cysteine removed by 3x5 min washing in 0.1M acetate buffer pH4 followed by 5 min in 50% ethanol. Microphotometry Microphotometric scanning measurements at the absorption maximum of 560 nm (DDDFB-staining) were performed as described in a previous paper [10]. Scanning of rectangular fields, each filled by a single stained cell, yielded both their mean optical density (OD) and their area (Ac;um2). The product of both (OD x Ac), the integrated extinction of one cell, is

50

G. Nohammer / Cytophotometric Investigations

proportional (e = 19000) to the moles of protein- bound 2-hydroxy-6-thionaphthol residues coupled with FB [13 ]. The mean integrated extinction (Etot) is the mean of all single integrated extinctions measured from single cells of a distinct sample. A minimum of 50 cells, distributed statistically, were scanned per sample. Chemicals 2,2'-dihydroxy-6-6'-dinaphthyldisulfide (DDD) and 2,4-dinitrofluorobenzene (2,4-DNFB) were obtained from Sigma Chemical Co., St. Louis, MO,USA; Fast blueB (FB), mercuric acetate, cysteine, 2-hydroxy-l-naphthaldehyde (HNA) and Merckoglas from Merck, Darmstadt, Germany.

3. Results and Discussion According to Torchinsky [11], metal salts (e.g.HgX2) should react in solution with disulfides (YSSY) forming 1.5 moles of mercaptide (YSHgX) and sulfinic acid (YSO2H). The proposed mechanism explaining the stoichiometry of the reaction was that first a complex is formed between a metal salt and a disulfide which is hydrolyzed the by the nucleophilic attack of OH" yielding one mole of mercatide and one mole of sulfenic acid (YSOH). Two moles of sulfenic acid should disproportionate then yielding one mole of sulfinic acid and one mole of thiol (YSH). YSH reacts with HgX2 to YSHgX +HX. Complex-formation between mercuric compounds and disulfides was reported by Brown and Edwards [14] and was used for the histochemical demonstration of protein disulfides [9]. Fixed cell preparations with Yoshida ascites tumor cells were used to investigate both the influence of the time of incubation of fixed cell preparations in mercuric acetate (HgAc) and of the time of the DDD-reaction running at pH4. We used pH4 for the reactions since at this pH DDD reacts exclusively with protein thiols (PSH) [10], and in addition protein mixed disulfides (PSSX) are not lost by disulfide exchange reactions with adjacent PSH [8]. The cell preparations were made by smearing one drop of the cell suspension in 0,15 M NaCl on slides, immediately fixing the distributed cells by spraying with ether ethanol (1:1), and storing the prefixed cell preparations in methanol (4°C). This procedure requires approximately 15 sec per slide. The cell preparations were indexed consecutively with a number according to the sequence of their fixation. The results of this experiment, illustrated in Fig. 1, was interpreted first as a complex relation between time used both for preincubations with HgAc and for DDD-reactions. However, the mean total extinction (Etot) values obtained microphotometrically after the Hg-catalyzed DDD-reaction (Hg//DDD) indicated that under the conditions used the histochemical reaction demonstrated protein disulfides (PSSP) and that DDD (RSSR) reacted to PSHgSR (R= 2hydroxy-6-thionaphthol). Much later, having more experience with Hg//DDD-reactions, we ranked the Etot values according to the index number of the cell preparations they were obtained from and we saw first that the results of this experiment might be influenced strongly by a biochemical oscillation, possibly of protein disulfides, which we intend to demonstrate (Fig. 2). To see if the Hg//DDD-reaction might be used for the histochemical demonstration of protein disulfides in tissue sections too, fresh frozen serial sections of rat liver were mounted on slides, fixed in methanol, and indexed with a number corresponding to the sequence in which the sections were cut. Fig. 3 illustrates that the parameters demonstrated histochemically depended on the sequence of cutting of the sections and that something was lost between 7 hours and 12 hours of the Hg//DDD-reaction. However, our results indicate that thiol-disulfide equilibria oscillate locally in tissues.

G. Nohammer I Cytophotometric Investigations

hour, of ODD (pH4)-raactkxi

Fig. 1. Influence of both time of the prelncubatlon of fixed Yoshida ascltes tumor cells with Hg// and the time of the ODD reaction at pH4. •,•,*,• symbolize the time of preincubation with HgAc at pH4 for 30 min, 2,11,and 24 hours, respectively. Bars symbolize the standard deviation of Etot values.

index number of preparation*

Fig. 2. Oscillations of protein disulfldes in Yoshida ascltes tumor cells. *,*,A,B symbolize the time of preincubation with HgAc at pH4 for 30 min, 2,11,and 24 hours, respectively. Bars symbolize the standard deviation of Etot values. The index numbers of the preparations correspond with the times of their fixation.

Index number of Mritt *

Fig. 3. Oscillations of protein disulfldes In rat liver serial sections. • and A symbolize mean optical densities measured after the Hg//DDD reaction at pH4 running 7 and 12 hours, respectively.

51

52

G. Nohammer / Cytophotometric Investigations

Fig. 4. Oscillations of both primary amlno groups and dlsurfides of proteins In Ehrilch ascltes tumor cells, y-axis: Etot values measured at 420 nm after HNA-staining and at 560 nm after Hg//DDD-FB-staining, respectively. A and • symbolize Etot values measured microphotometrically after HNA- and Hg//DDD-FB-staining, respectively.

Fig. 5. Dependence of the Hg//DDD reaction on both pH and Urns of the DDD-reactJon with EATC fixed at different times. •, •, A.X ,• symbolize pH 3,4.5,7. and 8.6 of the ODD reaction running after Hg//, respectively.

10

IS

20

2S

M

Hg. 6. Oscillations of the Etot values measured microphotometrically In EATC, fixed at different times, after both different times and pHs of the Hg//DDD reaction. •.•.A.x.t) symbolize pH 3,4,5,7,and 8.6,respectively, of the ODD reactions running after Hg//DDD. The index number of the preparations correspond with the time of their fixation.

G. Nohammer / Cytophotometric Investigations

53

To investigate whether other cellular parameters than protein disulfides oscillate too, cell preprations with EATC, fixed at different times, were used. The first series of 10 preparations was used for the histochemical demonstration of primary ammo groups of proteins using the HNA (2-hydroxy-l-naphthaldehyde) method [15], and the following series of 10 preparations was used for the Hg//DDD-reaction running at pH4 for 7 hours, demonstrating protein disulfides. Fig. 4 illustrates the oscillations of primary amino groups as well as of disulfides of cellular proteins. The dependence of the Etot-values measured after the Hg//DDD-reaction on the time of the DDD-reaction , demonstrated with rat liver serial sections (Fig. 3), indicated that the Hg//DDD-reaction in addition to PSHgSR might generate further products, and that the stability of these products might depend on both time and pH of the Hg//DDD-reaction. Therefore the dependence on both time and pH of the Hg//DDD-reaction was investigated using EATC. The cell preparations used were fixed at distinct times indicated by the sequence of their index numbers. The results of these investigations, illustrated in Fig. 5, indicate, that 1) a 5-day-DDD-reaction was necessary to obtain the highest Etot-values, 2) prolonged DDD-reactions running at pH3 to pH5 led to a considerable decrease of Etotvalues, 3) under the conditions of the Hg//DDD-reaction at pH7 only PSHgSR was generated and stable, and 4) the Hg//DDD-reaction at pH 8.6 exhibited the characteristics of a reaction transforming both sulfur atoms of PSSP generating PSHgSR and PSSR, both stable at pH 8,6. However, we used cell preparations not fixed synchronously. Later on, with the knowledge that the parameters demonstrated histochemically depend on the time of fixation, we ranked the Etot- values according to the index number of the cell preparations they were measured from. Again, the ranked Etot-values showed oscillations (Fig. 6), but the amplitudes of these oscillations compared with those observed after a short time Hg//DDD-reaction indicated the influence of at least one additional parameter, demonstrated using ODD at lower pH than pH7. To investigate the influence of prolonged times of incubation of fixed cell preparations in HgAc at pH4 on the time course of the Hg//DDD-reaction at pH4 we used EATC fixed at different times corresponding to the index number of the preparations. The results of this experiment, illustrated in Fig. 7, indicated that 1) the highest Etot values were obtained after 3 and 4 days of preincubation in HgAc pH4, 2) preincubation in HgAc pH4 for 1 day yielded the lowest Etot-values, and 3) reaction times with DDD at pH4 longer than 3 days led to a considerable decrease of Etot-values. Later on, with the knowledge of oscillations of the parameters demonstrated by the Hg//DDD-reaction, the Etot values illustrated in Fig. 7 were ranked again according to the index number of the cell preparations the Etot values were obtained from. The result of the ranking is illustrated in Fig. 8. The oscillating Etot values bring into question all of the three conclusions made above. They indicated no, or a much more complex influence of the time of preincubation in HgAc on the subsequent DDD-reactions. Finally we have learned that different cellular parameters oscillate, e.g. PSSP and PSSX. Therefore the following experiments were performed with EATC fixed at a distinct time according to the procedures described in methods, dependent on the type of experiment performed. The influence of both pH and time of Hg//DDD reaction was investigated with EATC, fixed synchronously, and preincubated in HgAc pH4 for 4 days, since we believed that this kind of preincubation should yield optimum results (Fig. 7). The results of this experiment, illustrated in Fig. 9, are: 1) After 1 day of Hg//DDD at pH3, the Etot value measured was much higher than the Etot values measured after Hg//DDD-reactions at pHvalues higher than 3, and the time course of the Hg//DDD-reaction at pH3 indicated a loss of a parameter during the first hours of the DDD-reaction, possibly an unstable or soluble substance under these conditions. 2) Compared with Hg//DDD reactions at pH higher than

54

G. Nohammer / Cytophotometric Investigations

Hg. 7. Influence of both the time of pralncubatlon of fixed preparation* with EATC In HgAc at pH4 and the time of the HgS/DDD reaction running at pH4. • A A.X symbolize 1,2,3,and 4 days.respectively, of preincubation in HgAc at pH4.

Hg. 8. Oscillation of the Etot values measured microphotometrlcally with EATC, fixed at different times, after both different time of preincubation with HgAc pH4 and different times of HgS/DDD reaction running at pH4. •.•.A.QO.A symbolize Etot values measured microphotometrically after 1,2,3,4,7, and 10 days, respectively, of Hg//DDD-reaction at pH4. Four series of 6 Etot values each (index numbers 1-6, 7-12, 13-18, 19-24, ) were measured with cell preparations preincubated with HgAc at pH4 for 1,2,3, and 4 days, respectively. The index numbers of the preparations correspond with the time of their fixation.

Hg. 9. Dependence of the Etot values measured microphotometrlcally with EATC, fixed synchronously, and prelncubated with HgAc at pH4 for 4 days on both pH and time of the Hg//DDD reaction. •JLA.x.*,*) symbolize Etot values measured after Hg//DDD reaction running at pH 3,4,5,6,7,and 8.6, respectively.

G. Nohamrner / Cytophotometric Investigations

55

pH4 (and with the exception of Hg//DDD at pH 8.6), the Etot values measured after 3 days of Hg//DDD at pH3 and pH4 were much higher. 3) The time- courses of the Hg//DDDreactions at pH 5,6,and 7 indicated that the parameters demonstrated needed 6 days of DDD-reaction. 4) The time- course of the Hg//DDD reaction at pH8.6 indicated a different mechanism of reaction compared to the Hg//DDD-reaction running at pH-values between pH3 and pH7. As shown in Fig. 5, the Etot-values measured after Hg//DDD at pH 8.6 indicated a reaction with both sulfur atoms of PSSP generating PSHgSR and PSSR, both stable at pH8.6 . The Hg//DDD-reaction could demonstrate both PSSP and PSH through formation of PSHgX [14]. Therefore we blocked irreversibly the PSH of cell preparations with EATC, fixed synchronously,using 2,4-dinitrofluorobenzene (2,4//), as described in methods. With cell preparations of the synchronously fixed EATC, used also for the experiment described before,which were pretreted with 2,4/7, we investigated again the influence of both pH and time of the Hg//DDD-reaction. The results of this experiment are illustrated in Fig. 10. Again, the Hg//DDD-reaction at pH3 yielded (except the Hg//DDD-reaction at pH8.6) the highest Etot values even after 2 days of DDD-reaction. A part of the products generated were lost with prolonged time of DDD-reaction at pH3. The Hg//DDD-reactions at pH 4,5,and 6 all yielded lower Etot values with a maximum obtained after 4 days of DDDreaction. The lowest Etot values were measured after Hg//DDD -reaction at pH6. The Hg//DDD-reaction at pH8.6 obviously ran according to another scheme, as discussed above. The last experiment showed that the Hg//DDD-reaction at pH3 yielded the highest and at pH6 the lowest Etot values. The hypothesis based on these results was that after Hg//DDD at pH3 (3 d), PSSP reacted to PSHgSR + PSOSR and PSSX (protein mixed disulfides) reacted to PSHgSR + XSOSR. Soluble low molecular weight XSOSR is lost from histochemical demonstration. PSOSR should be quite stable at pH3. Consequently, after Hg//DDD at pH6 (3d), PSSP as well as PSSX should have reacted to PSHgSR. PSOSR and XSOSR are unstable at pH6 and hydrolyzed to sulfmic acids which might be oxidized to sulfonic acids. To corroborate this hypothesis, the Hg//DDD-reactions at pH3 and pH6 were each investigated with EATC, of which a series of 8 sets of cell preparations were made, 2 preparations each fixed together at different times correlating with their index number. The time difference between the fixations was 34 sec. Subsequently, the cell preparations were blocked with 2,4/7 prior to Hg//DDD-reactions. Providing the following sequence of biochemical reactions is running, e.g. reduction of PSSP to PSH, generation of PSSX according to PSH + XSSX = PSSX + XSH (regeneration of XSH from XSSX), disulfide exchange according to PSH + PSSX = PSSP + XSH (regeneration of XSH from protein mixed disulfides), the Etot values measured after the 2,47/Hg//DDD-reaction at pH3 and pH6 should yield opposite oscillations. Fig. 11 illustrates opposite oscillations of the Etot values measured after the 2,4//Hg//DDD-reaction at pH3 and pH6 and shows that these oscillations might be impaired (index number 8) by other biochemical changes, possibly an increasing oxidation of PSH to PSOH. The Hg//DDD-reaction at pH3 demonstrates PSOSR too, generated from PSSP, which proved to be unstable at pH6. PSOSR should be formed by the reaction of PSOH, generated from PSSP by Hg//, reacting to PSHgX + PSOH (11). ODD (=RSSR) reacts with PSOH to PSOSR + RSH. The existence of stabilized forms of PSOH have been reported [16]. Therefore the existence of PSOH (and other DDD-reactive oxidized protein sulfur species) was investigated, using EATC fixed at different times. From a series of 15 sets of 4 cell preparations, each set fixed a distinct different time correlating with the index number of the cell preparations, the Etot values were measured after Hg//DDD-reactions at pH3 and pH6 as well as after 2,4//Hg//DDD reactions at pH3 and pH6.

56

G. Nohammer / Cytophotometric Investigations

Rg. 10. Dependence of the Etot values measured mterophotometrlcally with EATC, fixed synchronously, pratraated with 2,4/7, and preincubated with HgAc at pH4 for 4 days on both pH and time of the Ho/TDDD reaction. *,B,A,x .*,• symbolize Etot values measured after 2,4//Hg//DDD-reaction running at pH 3,4,5,6,7, and 8,6, respectively.

Rg. 11. Oscillating conversion of protein dlsulflde* and protein mixed disuhldes In EATC, fixed at different times, and pretreated with 2,4/7. • and • symbolize Etot values measured microphotometrically after 2,4//Hg//DDD reaction running at pH 3 and pH6, respectively.

Rg. 12. Oscillations of Etot values measured microphotometrically with EATC, fixed at different times, after 3 days of Hg/7DDD reaction and 2,477Hg/7DDO reaction at pH3 and pH6, respectively. A.B.A.n symbolize Etot values measured after Hg//DDD at pH3. Hg//DDD at pH6, 2,4//Hg//DDD at pH3, and 2,4//Hg//DDD at pH6 after 3 days of ODD reaction, respectively. The index numbers of the preparations correspond with the time of fixation.

G. Nohammer / Cytophotometric Investigations

57

Figure 12 illustrates oscillations of the parameters demonstrated by the four different histochemical methods used. The differences between the Etot values measured after Hg//DDD reactions at pH3 and pH6, illustrated in Fig. 13, should be generated by parameters demonstrated histochemically with DDD and stable at pH3 but not at pH6, e.g. PSOSR and possibly PSOaSR. The differences between the Etot values measured after 2,4//Hg//DDD reactions at pH3 and pH6, illustrated in fig 13, should be generated by parameters, not destroyed by 2,4/7, demonstrated histochemically with DDD, and stable at pH3 but not at pH6, e.g. PSOSR generated from PSSP. PSSP react to PSHgSR (stable at pH6) and PSOSR (quite stable at pH3). To see if PSOSR generated by the Hg//DDD reaction at pH3 from PSSP is the only product formed with DDD and stable at pH3, the difference-Etot values of the differences between the Etot values measured after the Hg//DDD reactions at pH3 and pH6, and 2,4//Hg//DDD-reactions at pH3 and pH6, were calculated, illustrated in Fig. 13. These double-difference-values, oscillating too, should be due to DDD-reactive protein species, not destroyed by Hg//, and stable at pH3, possibly PSOSR, generated by the reaction of DDD at pH3 with stabilized PSOH [16].

Fig. 13. Oscillating conversion of protein dlsulfldes, protein mixed dlsulffdes and oxidized protein sulfur in EATC, fixed at different times.

• symbolize the difference calculated

between the Etot values measured after the Hg//DDD reaction running at pH3 and pH6.

•

symbolize the difference calculated between the Etot values measured after the 2,4//Hg//DDD reaction running at pH3 and pH6. A symbolize the difference calculated between the difference Etot values symbolized by • and • .

The products generated by the Hg//DDD reaction and shown to stable at pH 3, e.g. PSOSR, are the result of the Hg/TDDD reaction with PSSP as well as of DDD reactions with other oxidized protein sulfur species, as shown in Fig. 13. To exclude the possibility that these other oxidized protein sulfur species are produced during the preincubation of cell preparations with HgAc, cell preparations were made with EATC, fixed synchronously. With these preparations the dependence of the DDD reaction (not catalyzed by HgAc) on both pH and time was investigated. The results, illustrated in Fig. 14, were:l) The DDD reaction at pH 3 and pH2.5 generated products which were quite stable at this low pH, unstable even at pH4 and destroyed at pH 6 of the DDD reaction. Preincubation of the fixed cell preparations in 0.2 M cysteine at pH4 for 2 hours led to a considerable decrease of the Etot values measured after DDD reactions at all pH values investigated. This decrease of the Etot values was pronounced when cell preparations were preincubated in 0.2 M cysteine at pH4 for 6 hours. Preincubation of fixed cell preparations with 2,4/7 abolished all protein species reactive with DDD at pH values between 2.5 and 6.

58

G. Nohammer / Cytophotometric Investigations 120 i

E.

Fig. 14. Dependence of the Etot values measured mtorophotometrteaily with EATC, fixed synchronously, after DDD-FB-stalnlng on pH and time of the ODD reaction and on preincubation of the cell preparations with cystalne and 2,4/7. •,•,*.• (full) symbolize the Etot values measured microphotometrically with EATC, fixed synchronously.after the ODD reaction running at pH 2.5,3,4, and 6, respectively. <>O^,O (open) symbolize the Etot values measured microphotometrically with EATC, fixed synchronously and preincubated with 0.2M cysteine at pH4 for 2 hours prior to the ODD reaction running at pH 2.5, 3,4,and 6. respectively. (-), (+) symbolize Etot values measured microphotometrically with EATC, fixed synchronously, and pretreated with 2,4// prior to the DDD reaction running at pH 2.5 and pH6, respectively, x (dotted line) and * (broken line) symbolize Etot values measured microphotometrically with EATC, fixed synchronously, and preincubated with 0.2M cysteine at pH4 for 6 hours prior to DDDreaction running at pH 2.5 and pH6, respectively.

Summarizing, the DDD reaction performed at low pH revealed the existence of oxidized protein sulfur species which were reduced even under mild reducing conditions, e.g. 0.2 M cysteine at pH4, and were destroyed by 2,4/7 and lost at pH values higher than 4 of the DDD reaction. Such oxidized protein sulfur species could be stabilized protein sulfenic acids (PSOH), methionine sulfoxide (PSOCHs), oxidized protein disulfides (PSOSP)(17), and mixed disulfides (PSOSX), and possibly other forms of oxidized protein sulfur. Under the conditions of the Hg//DDD reaction, protein disulfides (PSSP) react to PSHgSR and PSOSR with DDD (=RSSR) and protein mixed disulfides to PSHgSR and XSOSR. PSHgSR was shown to be stable even at high pH, PSOSR and XSOSR only at low pH. Soluble low molecular weight XSOSR is lost from histochemical demonstration. At least some of the oxidized protein sulfur species that reacted with DDD and proved to be stable at low pH were not destroyed by Hg//. We could show that all the cellular protein parameters demonstrated after Hg//DDD reactions oscillate, revealing the time course of their generation and biochemical transformation, if demonstrated using cell preparations fixed at different times.

4. Summary In order to use mercuric salts, reported to interact with disulfides, for the histochemical demonstration of protein disulfides, investigations were performed on the dependence of the reactions using 2,2'-dihydroxy-6,6'-dinaphthyl-disulfide (DDD) with fixed cell preparations preincubated with mercuric acetate (HgAc), on pH and time of DDD-reaction as well as on the time of preincubation with HgAc. The first studies, performed with Ehrlich ascites tumor cells (EATC) and Yoshida ascites tumor cells (YOATC), were performed with cell preparations, fixed at different

G. Ndhammer / Cytophotometric Investigations

59

times, corresponding to the index number of the preparations. The results, obtained microphotometrically, of these first investigations, indicating a Hg-catalyzed DDD-reaction with protein disulfides, were somehow confusing and led to some wrong conclusions. However, ranking of the results according to the time of fixation of the cell preparations they were obtained from revealed biochemical oscillations of the parameters demonstrated histochemically. Oscillations of protein disulfides were demonstrated with rat liver serial sections too. With the knowledge that the parameters to be demonstrated histochemically oscillate, further investigations were performed with preparations of EATC fixed synchronously at distinct times. Investigations on the influence of both time and pH of the Hg-catalyzed DDD-reaction performed with cell preparations, either fixed only, or pretreated using 2,4-dinitrofluoro-benzene (2,4/7), revealed that 1) protein disulfides (PSSP) were transformed to PSHgSR and PSOSR (RSSR=DDD), PSOSR stable only at low pH, 2) protein mixed disulfides (PSSX) were transformed to PSHgSR and XSOSR, the latter lost for histochemical demonstration, 3) also protein sulfur species other than disulfides reacted with DDD (also without Hg-catalysis). These were stable only at low pH and lost for histochemical demonstration by preincubation either with cysteine or with 2,4/7. The chemical properties of these species indicated the existence of oxidized protein sulfur species, e.g. PSOH, PSOSP, PSOCH3 and possibly others, that reacted with DDD to PSOSR, and possibly to PSOjSR, both stable only at low pH. Protein disulfides , protein mixed disulfides, and oxidized protein sulfur species were shown to oscillate.

Acknowledgments The author thanks Monika Reiter-Khabir, Doris Celotto, and Gisela Pongratz for technical assistance. Critical reading of the manuscript by Dr. J. Ross Stevenson is gratefully acknowledged.

References 1. Gilbert, H.F. Redox control of enzyme activities by thiol/disulfide exchange. Methods Enzymol., 107: 330351, 1984. 2. Powis,G,,Briehl, M., and Oblong,J. Redox signalling and the control of cell growth and death. Pharmac. Ther., 68:149-173, 1995. 3. Storz, G., and Polla, B.S. Transcriptional regulators of oxidative stress-inducible genes in prokaryotes and eukaryotes. In: Feige, U., Morimoto, R.I., Yahara, I., and Polla, B. (Eds.), Stress-inducible Cellular Responses, Birkhauser Verlag, Basel, Switzerland, pp. 239-254, 1996. 4. Slater, T.F., Bajardi, F., Benedetto, C., Bussolati, G., Cianfano, S., Dianzani, M.U., Ghiringhello, B., Nohammer, G., Rojanapo, W., and Schauenstein, E. Protein thiols in normal and neoplastic human uterine cervix. FEBS Letters, 187:267-271,1985. 5. Ndhammer, G., Bajardi, F., Benedetto, C., Schauenstein, E., and Slater, T.F. Quantitative cytospectrophotometric studies on protein thiols and reactive protein disulfides in samples of normal human uterine cervix and on samples obtained from patients with dysplasia and carcinoma-in-situ. Br. J. Cancer, 53:217-222, 1986. 6. Benedetto, C., Bajardi, F., Ghiringhello, B., Marozio, L, Ndhammer, G., Phitakpraiwan, P., Rojanapo, W., Schauenstein, E., and Slater, T.F. Quantitative measurements of the changes in protein thiols in cervical intraepithelial neoplasia and in carcinoma of the human uterine cervix provide evidence of a biochemical field effect. Cancer Res. 50:6663-6667,1990. 7. Nohammer, G., Bajardi, F., Benedetto, C., Kresbach, H., Rojanapo, W., Schauenstein, E., and Slater, T.F. Histophotometric quantification of the field effect and the extended field effect of tumors. Free Rad. Res. Comms., 7:129-137, 1989. 8. Nohammer, G., Desoye, G., and Khoschsorur, G. Quantitative cytospectrophotometrical determination of total protein thiols with «Mercurochrom». Optimization and calibration of the histochemical reaction. Histochemistry, 71:291-300, 1981.

60

G. Ndhammer / Cytophotometric Investigations

9. Ndhammer, G., and Desoye, G. Mercurochrom can be used for the histochemical demonstration and microphotometric quantification of both protein thiols and protein(mixed)disulfides. Histochem. CeH Bid., 107:383-390.1997. 10. Ndhammer, G. Quantitative microspectroprtotornetrical determination of protein thiols and disulfides with 2,2'-dihydroxy-6,6'-dinaphthyldisulfide (ODD). Histochemistry, 75:219-250,1982. 11. Torchinsky, Yu.M. Sulfur in proteins. Pergamon press, Oxford, England, pp. 74-75,1981. 12. Barmett, J.R. and Seligman.A.M. Histochemical demonstration of protein-bound SH-groups. Science, 116: 323-327,1952. 13. Esterbauer.H. Beitrag zum quantitattven histochemischen Nachweis von Sulfnydrytgruppen mil der DDDFarbung. I. Untersuchung der Farbstoffe. Acta histochem., 42:351-355,1972. 14. Brown.P.R., and Edwards, J.O. Reaction of disulfides with mercuric ions. Biochemistry, 8: 1200-1202, 1969. 15. Ndhammer, G. Histochemical demonstration of primary amino groups with 2-hydroxy-1-naphthaldehyde (HNA): Optimization of the method. Acta histochem., 86:167-176,1989. 16.16.Claibome, A., Miller, H., Parsonage, D., and Ross, R.P. Protein-sulfenic acid stabilization and function in enzyme catalysis and gene regulation. FASEB J., 7:1483-1490,1993. 17. Block, E., Gulati, H., Putman, D., Sha, D., You, N., and Zhao, S.-H. Allium chemistry: Synthesis of 1[Alk(en)ylsulfinyl]propyl alk(en)yldisulfides (Cepaenes), antithrombotic flavorants from homogenates of onion (allium cepa). J. Agric. Food Chem., 45:4414-4422,1997.

61

Thiol Metabolism and Redox Regulation of Cellular Functions A. Pompella et al. (Eds.) 1OS Press, 2002

Protection by Pantothenic Acid against Apoptosis and Cell Damage by Oxygen Free Radicals - The Role of Glutathione Lech WOJTCZAK and Vyacheslav S. SLYSHENKOV* Nencki Institute of Experimental Biology, Pasteura 3, 02-093 Warsaw, Poland; e-mail: [email protected] * Present address: Institute of Biochemistry, The National Academy of Sciences of Belarus, Grodno, Belarus

1. Introduction Pantothenic acid belongs to vitamin B group (it is usually termed vitamin Bs) and is the building stone of coenzyme A (Formula 1). It is not synthesised by animal tissues, but is produced by intestinal bacteria and is ubiquitous in food products. Therefore, pantothenic acid deficiency in animal and human subjects is not known, except for experimental conditions. Nevertheless beneficial effects of pantothenic acid and its reduced derivative, pantothenol, have been observed in radiation injury [ 1 -4] and wound healing, especially in tissues exposed to atmospheric oxygen, like skin [5-10], lung epithelium [11] and eye cornea [12,13].

< \

Pantothenic acid moiety

fi C.

H,C — o — pP — o —

IIc p

"371

Y '

o //

0-CH-C -C -C.

i OH i Y|

O //

H3C

H

N —H I CH,

Formula 1. Coenzyme A

Protection by pantothenic acid against heart and liver injury caused by ischemia and oxidative stress has also been reported [14-20]. Thus, beneficial effects of pantothenic acid and pantothenol have been reported under conditions when damage by oxygen free radicals might be suspected. This report will briefly describe research carried out in the authors' laboratories on the protective action of pantothenic acid and its derivatives on mammalian cells subjected to various kinds of oxidative stress. It will also present the mechanism underlying these effects that involves glutathione and, possibly, other cellular thiols.

62

L Wojtczak and V.S. Slyshenkov I Protection by Pantothenic Acid

2. Protection by Pantothenic Acid against Oxidative Damage of Ehrlich Ascites Tumour Cells Ehrlich ascites tumour cells extracted from the peritoneal cavity of mouse proved to be a convenient model. When subjected to reactive oxygen species (ROS) generated either chemically or by ultraviolet (UV) irradiation, they manifest typical symptoms of oxidative stress, like accumulation of lipid peroxidation products, damage to the plasma membrane that results in its leakiness, and impaired mitochondrial ATP synthesis [21,22]. When lipid peroxidation was measured as accumulation of thiobarbituric acid-reactive compounds (expressed as malondialdehyde) and plasma membrane damage by leak of lactate dehydrogenase, a parallelism of both processes could be observed (Fig. 1). Preincubation of the cells with pantothenic acid or pantothenol partly protected against lipid peroxidation induced by ROS. This protection was concentration-dependent and was accompanied by a partial protection against damage to the plasma membrane. As shown (Fig. 2), pantothenic acid, pantothenol and pantethine (a thiol-containing derivative of pantothenic acid) were equally active, whereas homopantothenic acid exhibited a low protective effect.

0

20

40

60

Irradiation time (min)

Fig. 1. Lipid peroxidation and permeabilization of the plasma membrane induced by UV irradiation in Ehrlich ascites tumour cells. •, Accumulation of thiobarbituric acid-reactive compounds (determined as malondialdehyde); O, leakage of lactate dehydrogenase (LOH, expressed as percentage of total lactate dehydrogenase liberated after solubilisation of the cells with 0.2% digitonin). From [22].

The protective effect of pantothenic acid and its derivatives against lipid peroxidation and permeabilization of the plasma membrane required a preincubation of the cells with these compounds for at least several minutes. Moreover, a substantial reduction of lipid peroxidation and diminution of the plasma membrane leakiness could be obtained only after preincubation at room temperature or at 32°C. Preincubation at 0°C for up to 60 min had no effect on either lipid peroxidation or plasma membrane damage (Fig.3).

63

L Wojtcz.uk and V.S. Slyshenkov I Protection by Pantothenic Acid

Fig. 2. Effect of pantothenic acid and related compounds on lipid peroxidation and damage of the plasma membrane in Ehrlich ascites tumour cells. The cells were preincubated with 1 mM pantothenic acid or its derivatives for 40 min at 22°C and then incubated with the Fenton reagent (0.4 mM FeCI2 + 0.2 mM H2O2). Formation of malondialdehyde was measured after 10 min incubation with the Fenton reagent and the leak of lactate dehydrogenase (LDH) was determined after 60 min. The leak is expressed as percentage of the total activity of lactate dehydrogenase liberated after complete permeabilization of the cells with digitonin. a, Before the Fenton reaction; b - f, after the Fenton reaction; b, the cells preincubated without additions; c, preincubated with pantothenic acid; d, preincubated with pantothenol; e, preincubated with pantethine; f, preincubated with homopantothenic acid. Data taken from [21].

If f2 I Q-

A

50

II

* 25

0 o

1i 20

40

Time (min)

60

20

40

60

Time (min)

Fig. 3. Protective effect of pantothenic acid against lipid peroxidation and plasma membrane damage by oxygen free radicals in Ehrlich ascites tumour cells. The cells were preincubated with 1 mM pantothenic acid at 0°C (O), 22°C (•) and 32°C (A) for the time indicated at the abscissa, collected by centrifugation and then incubated in the same medium at 22°C for 10 min with 0.4 mM FeCI2 + 0.2 mM H2O2. Lipid peroxidation (A) was measured by accumulation of thiobarbituric acid-reactive compounds (expressed as malondialdehyde) and plasma membrane damage (B) as leakage of lactate dehydrogenase. Partly from [21].

64

L Wojtcz&k and V.S. Slyshenkov I Protection by Pantothenic Acid

ADP

CCCP

ADP

CCCP

ADP

CCCP

Rg. 4. Effect of UV irradiation on membrane potential of Ehrlich ascftes mitochondria. The celts were irradiated for various periods of time, followed by permeabilization with dkjrtonin, and the mitochondria! membrane potential (Ay) was measured using the safranine O fluorescence assay. Energization was induced by addition of succinate (Succ.). A complete collapse of Ay was obtained by addition of carbonyl cyanide m-chlorophenylhydrazone (CCCP). Indications, as exemplified in traces A and B, are: Ay, maximum membrane potential of fully energized mitochondria; dAy, difference between membrane potentials at State 4 and State 3; Ay/f, initial rate of Ay formation; f, time required for Ay to return to its initial value after addition of ADP. The irradiation time was 0, 5,10 and 20 min for traces A, B, C and D, respectively. From [22].

These results indicated that pantothenic acid or pantothenol did not behave as free radical scavengers but, rather, exerted a metabolic effect. In fact, these compounds did not protect phospholipids in abiotic system (in form of liposomes) against lipid peroxidation by oxygen free radicals generated by the Fenton reaction (not shown). ROS not only impaired the cell surface but also exerted a deeper action on Ehrlich ascites cells by damaging their energy generating system of mitochondria. This is illustrated in Fig. 4 where building of the mitochondrial membrane potential (Ay) in Ehrlich ascites tumour cells subjected to ultraviolet irradiation is depicted. As shown, irradiation slows down formation of A\y and decreases the rate of ADP phosphorylation. ROS generated by UV irradiation (Fig. 5) or the Fenton reaction (not shown) increased the rate of resting state respiration but decreased the active state respiration and decreased the rate and the efficiency of ATP synthesis. All these parameters were normalised when the attack by ROS was preceded by preincubation of the cells with pantothenic acid (Table 1). Pantothenic acid is a precursor of CoA. It can be therefore inferred that its beneficial effect in various kinds of cell damage by ROS is related to the increased content or stimulated biosynthesis of this coenzyme. In fact, we found that the content of CoA in Ehrlich ascites cells briefly incubated with another ROS generator, terf-butyl hydroperoxide, decreased almost to zero, but if the cells were preincubated with pantothenol, the content of CoA was doubled and even in the presence of terf-butyl hydroperoxide it was slightly increased and not decreased (Fig. 6).

65

L. Wojtczak and V.S. Slyshenkov / Protection by Pantothenic Add

C Q. 0) _

4.0

20

c "3 +•> o

15

25

3.5

f-4—t

£ .E (0 0)

I**

o S.

3.0

•.01

O 2.5 51 < 2'° 1.5

10

0).£ >» c x •£ O o o E c

-I o o a> E toS

1.0 10

0.5

Irradiation time (min)

8

10

Irradiation time (min)

Fig. 5. Effect of UV irradiation on energy coupling parameters of Ehrlich ascites tumour mitochondria. Intact cells were irradiated for the time indicated at the abscissa, permeabilized with digitonin and their respiration was subsequently measured with succinate as substrate. Active respiration (State 3) was induced by adding 5 mM ADP, and the uncoupled state by 0.1 mM 2,4dinitrophenol. ADP/O ratio was calculated from the amount of O2 consumed to phosphorylate a known amount of ADP as indicated by transition to the resting state (State 4). The rate of phosphorylation was estimated from the time elapsed between addition of ADP and transition to State 4. Indications: •, State 3; O, State 4; V, uncoupled state; •, ADP/O ratio; D, rate of ADP phosphorylation. From [22].

Table 1. Protective effect of pantothenic acid against mitochondria damage by oxygen free radicals.

Control

+ Pantothenic acid

Resting state respiration (State 4)

143 ±9

105 ±4

Active state respiration (State 3)

64 ±9

99 ±7

Respiratory control ratio

44 ±3

93 ±4

Rate of ATP synthesis

47 ±7

92 ±7

ADP/O ratio

71 ±9

94 ±6

Membrane potential in State 4

87 ±3

101 ±4

Ehrlich ascites tumour cells were preincubated 40 min without or with 1 mM pantothenic acid, followed by incubation with 0.2 mM Fe2* + 0.1 mM H2O2 (Fenton reaction) for 5 min. Mitochondria! energy coupling parameters were measured in cells permeabilized with digitonin. All values are expressed as percentage of the respective parameters in cells not subjected to oxidative stress. From [22].

It was subsequently found [22] that not only CoA but also glutathione contents were substantially increased by pantothenic acid. Pantothenic acid also protected against the decrease of glutathione and its oxidation during irradiation of the cells with ultraviolet light (Table 2).

66

L Wojtczak and V.S. Slyshenkov I Protection by Pantothenic Acid

Fig. 6. Effect of oxidative stress on the content of CoA in Ehriich ascites tumour cells and the protective action of pantothenic acid and pantothenol. The cells were preincubated with 1 mM pantothenic acid or pantothenol at 22°C for 40 min. Thereafter, the cells were collected by centrifugation and incubated at the same temperature for 40 min without (white columns) or with (dashed columns) 10 mM tert-butyl hydroperoxide. A, not preincubated control; B, preincubation without CoA precursor C, preincubation with pantothenic acid; D, preincubation with pantothenol. From [21].

Table 2. Effect of preincubation with pantothenic acid and UV irradiation on the content of glutathione in Ehriich ascites tumour cells. Treatment

GSH

GSSG

(nmol/mg protein)

(nmol/mg protein)

before irradiation

16.6 ± 1.9 (5)

0.38 ± 0.08 (2)

after irradiation

10.6 ±2.6 (3)"

1.10 ±0.40 (2)

Preincubation without pantothenic acid

Preincubation with pantothenic acid before irradiation

23.2 ± 1.2 (5)b

0.38 ± 0.02 (2)

after irradiation

18.5±3.0(3)c

0.70 ±0.10(2)

The cells were preincubated for 40 min at 22°C without or with 1 mM pantothenic acid and thereafter irradiated with ultraviolet light during 10 min. For GSH the data are means ± SD for the number of experiments indicated in parentheses; for GSSG mean values ± range for two experiments are shown. Statistical significance, for GSH only (Student's paired t-test):' p < 0.02 with respect to non-irradiated cells; " p< 0.001 with respect to cells preincubated without pantothenic acid;c p < 0.02 with respect to irradiated cells preincubated without pantothenic acid. From [22].

3. Pantothenic Acid Protects against Free Radical-induced Apoptosis Such profound damage of vital cell functions produced by prolonged treatment of the cells with high doses of ROS irrevocably leads to cell necrosis. However, low doses of ROS or a brief exposure to their action result in the programmed cell death, the so-called apoptosis. This kind of cell death, in contrast to necrosis, facilitates the removal of cell debris by

L. Wojtczak and V.S. Slyshenkov I Protection by Pantothenic Acid

67

macrophages and is not accompanied by inflammation of adjacent tissues. Apoptosis is a complex process that can be elicited by numerous external and internal stimuli, among them oxygen free radicals, and is controlled by several factors (for reviews see [23-25]). Because Ehrlich ascites cells, when seriously damaged, undergo a necrotic decay rather than apoptosis, the experiments on the effect of pantothenic acid and its derivatives on apoptosis were performed with human leukemic lymphocytes (Jurkat cells) [26]. These cells can be induced to apoptosis be brief irradiation with ultraviolet light. This was accompanied by lipid peroxidation and a drastic decrease of the glutathione content. Preincubation of the cells with pantothenic acid increased their glutathione content by more than 50% and increased the GSH/GSSG ratio by approximately the same factor (Fig. 7). A similar effect was exerted by N-acetylcysteine, the immediate precursor of glutathione. Interestingly, the effect of N-acetylcysteine was expressed already after 1 h, whereas that of pantothenic acid was fully manifested only after 3 h, indicating that the latter effect was more complex. Such preincubation diminished lipid peroxidation after UV irradiation (Fig. 8) and alleviated the decrease of glutathione (Fig. 9).

5 o

Fig. 7. Effect of pantothenic acid and N-acetylcysteine on the content and the redox state of glutathione in Jurkat cells. The cells were incubated with 1 mM pantothenic acid (•) or 5 mM Nacetylcysteine (O) and analysed for total gl.utathione (A) and the GSH/GSSG ratio (B) at various times of incubation. The results are expressed in percentage of the values in cells incubated without additions which amounted to 15.1 ± 1.8 nmol for total glutathione per mg protein and to 19.9 ± 1.1 for the GSH/GSSG ratio. The points are mean values ± SD for at least 3 experiments. From [26] modified.

UV irradiation as weak as 50 joule/m resulted in the appearance of about 25% of apoptotic cells after 4 - 6 h. Irradiation at the energy output of 100 and 150 joule/m produced about 60% of apoptosis (Fig. 10). Preincubation of the cells with pantothenic acid before irradiation prevented the apoptosis in a dose-dependent way, a significant protection being observed at 1 mM concentration. A partial protection against UV-induced apoptosis was also observed after preincubation of the cells with N-acetylcysteine. Interestingly, 1 mM pantothenic acid appeared to be a better protector than 5 mM N-acetylcysteine (Fig. 10).

68

L Wojtczak and V.S. Slyshenkov I Protection by Pantothenic Acid

Incubation time (h)

Incubation time (h)

Rg. 8. Protection by pantothenic acid and N-acetylcysteine against UV-induced lipid peroxidation as expressed by accumulation of conjugated dienes. Jurkat cells were preincubated for 3 h with 1 mM pantothenic acid (•), 5 mM N-acetylcysteine (O), or without additions (D), irradiated with UV light at the energy output of 100 or 150 j/m2, and analysed for conjugated dienes at various times following the irradiation. The results are expressed in percentage of the value before irradiation. From [26].

25

(•

~~Q

25

0 I

.

0

2

1

1

'

4

6

(

Incubation time (h)

•

0

6 7

i

1

t Incubation Mm* (h)

Fig. 9. Effect of UV irradiation on the content of total glutathione (A) and its redox state (B). Jurkat cells, preincubated for 3 h with pantothenic acid or N-acetylcysteine as in Rg. 7, were irradiation with UV light at the energy output of 100 j/m2 and then further incubated in the presence or absence of the same additions. The values for 0 time in this Figure are those for 3 h in Fig. 7. Description of the symbols is the same as in Fig. 8. From [26], modified.

Rg. 10. Effect of pantothenic acid and N-acetylcysteine on UV-induced apoptosis. Jurkat cells were preincubated for 3 h in the absence or presence of pantothenic acid or N-acetylcysteine, irradiated with UV at the energy output as indicated, and post-incubated with the same additions for 6 h. After that time apoptosis was determined using Hoechst staining. Indications: white columns, cells preincubated without additions; black columns, cells preincubated with 1 mM pantothenic acid; dashed columns, cells preincubated with 5 mM N-acetylcysteine. The proportion of apoptotic cells in non-irradiated controls was less than 2%. From [26].

;

L Wojtczak and V.S. Slyshenkov I Protection by Pantothenic Acid

69

4. Protection against Radiation Damage of Organs in Whole Animals These results obtained with cells in vitro as well as literature reports on the beneficial effects of pantothenic acid and its derivatives in various pathologies [1-20] prompted us to investigate whether these compound may protect whole animals against low doses of yradiation. One of the mechanisms of the damaging effects of ionising radiation is the generation of ROS [27,28]. Using adult rats, we concentrated our study on the liver, because this organ, characterised by its intense metabolism, may be especially prone to radiationevoked metabolic injury. The animals were exposed once a week to brief irradiation from a cobalt bomb, receiving a total dose of 0.75 Gy during three weeks. The animals were killed 1 h, 24 h and 7 days following the last irradiation and their livers were analysed. Experimental animals obtained pantothenol (26 mg/kg body weight) during two days before each irradiation, whereas the control animals were not supplemented. It was found [29,30] that such low-dose irradiation resulted in a considerable peroxidation of liver lipids, manifested by accumulation of peroxidation products, conjugated dienes and malondialdehyde (plus other thiobarbituric acid-reactive compounds), during the first hour following the final exposure. These peroxidation products were, however, partly eliminated from the liver already during the following 24 h and completely disappeared after 7 days. Parallel to this, liver glutathione and CoA contents decreased 1 h following the irradiation by about 25% and did not completely recover even after 7 days. The most drastic change was found for the GSH/GSSG ratio, the value of which decreased to about 15% of the non-irradiated control. Supplementation with pantothenol completely protected against all these peroxidative changes (Fig. 11). y-Irradiation also decreased the activities of some enzymes involved in preventing oxidative stress, like catalase, glutathione peroxidase, glutathione reductase and the so-called malic enzyme that is involved in keeping NADP in the reduced state. These effects were also prevented by pantothenol feeding (not shown).

5. Oxidative Stress, Apoptosis and Glutathione Apoptosis is a common process in multicellular organisms. It enables to eliminate single cells and their assemblies when their natural biological functions have been terminated, e.g. during morphogenesis and embryogenesis, or when they became damaged or mutated. The mechanism of apoptosis involves two partly interdependent routes, one of them initiated by stimulation of 'death receptors' at the cell surface, the other one involving mitochondria [24,25]. The 'mitochondrial pathway' is a multistep and extremely complex process. It includes, among other factors, pro-apoptotic proteins Bax and Bid that associate with the mitochondrial surface and promote the release of cytochrome c and of another proteinaceous apoptosis-inducing factor AIF [24,31]. The two latter proteins activate intracellular proteases, caspases, that initiate self-digestion of the cell and fragmentation of its DNA (Fig. 12). The action of pro-apoptotic proteins, Bax and Bid, is counteracted by anti-apoptotic protein Bcl-2. The mechanism by which Bax and Bid induce cytochrome c release from mitochondria is still debatable. It is generally agreed upon that an important role is played by the so-called permeability transition pore, a non-selective channel that opens in the inner mitochondrial membrane under certain conditions [33,34]. This pore, being presumably associated with the contact sites between the outer and the inner mitochondrial membranes [35], is regulated, among others, by the redox state of cell thiols, including glutathione, and is sensitive to oxidative stress [36,37]. In fact, among factors that trigger the programmed cell

L Wojtczak and V.S. Slyshenkov I Protection by Pantothenic Acid

70

death are oxygen free radicals [38-40]. Their action may involve promotion of the open state of the permeability transition pore [36,37] and binding of pro-apoptotic proteins with the mitochondrial surface (Fig. 12). Numerous reports point to oxygen free radicals and oxidative stress as factors inducing pore opening and to reduced glutathione as promoting its closure. Conjugated di- and trienes

Malondialdehyde

~ 100

1h

24h

7d

1h

24h

7d

Total glutathione

CoA

1h

24h

76

1h

24h

7d

NAD/NADH

GSH/GSSG 140 120 100.

1h

24h

7d

Fig. 11. Changes of several biochemical parameters in livers of ^irradiated rats and the effect of pantothenol. White columns, irradiated animals without further treatment; black columns, irradiated animals supplemented with pantothenol. From [30], modified.

The same factors also favour apoptosis or prevent it, respectively [41-46]. However, the permeability transition pore alone is too small to permeate cytochrome c (molecular mass 12.5 kDa). It has been therefore proposed [47] that the multiprotein assembly of the contact site, together with the permeability pore, can be triggered by the pro-apoptotic proteins Bax and Bid to release cytochrome c from the mitochondrion (Fig. 12). In fact, recent data indicate that Bid [48] and Bax (Wieckowski and Wojtczak, unpublished) preferentially associate with the contact sites and thus may promote liberation of cytochrome c into the cytosol. It is also worthy to note that cardiolipin, the unique phospholipid of the inner mitochondrial membrane.

L. Wojtczak and V.S. Stvshenkov I Protection by Pantothenic Acid

1\

has been found to be a specific target for Bid [49]. Cardiolipin is characterised by its high content of polyunsaturated fatty acids that are known to be extremely sensitive to peroxiodation.

Fig. 12. Simplified scheme of the mitochondria! pathway of apoptosis. The pathway is triggered by various "death signals", as oxygen reactive species (ROS), DNA damage etc., that promote binding of the pro-apoptotic protein Bax to the outer mitochondrial membrane, most likely at the contact sites between the two membranes, and its association with the permeability transition pore (FTP). This enables the release of cytochrome c (•) and the apoptosis-inducing factor (AIF, •) from the intermembrane compartment to the cytosol. An elevated intramitochondrial Ca2* level and ROS production facilitate this process by promoting FTP opening. Once in the cytosol, cytochrome c and AIF, in co-operation with a cytosolic factor, Apaf-1 (not indicated), activate caspase-9 and subsequently other members of the caspase family, thus initiating self-digestion of the cell and nuclear DNA fragmentation, eventually leading to apoptotic cell death. Association of Bax with mitochondria is prevented by the anti-apoptotic protein Bcl-2. ROS can be decomposed by Mncontaining (mitochondrial) and Cu.Zn-containing (cytosolic) superoxide dismutases (SOD), catalase and glutathione peroxidase (GPx). Stimulation of ROS production is exemplified here by UV and ionising radiation and by two anticancer drugs, adriamycin (extramitochondrially) and c;s-1-hydroxy4-(1-naphthyl)-6-octylpiperidine-2-one (BMD188, intramitochondrially). Activation is indicated as (+), and inhibition as (-). From [32].

Hence, glutathione is an essential element controlling apoptosis and other kinds of ROS-induced cytotoxicity [41-46]. Because all protective effects of pantothenic acid and its derivatives described here were correlated with the increased level of glutathione, it seems justified to propose that these compounds exert their beneficial action by promoting the synthesis of glutathione or preventing its degradation and/or efflux from the cell. It could be

72

L Wojtczak and V.S. Slyshenkov I Protection by Pantothenic Acid

expected that pantothenic acid, as precursor and building stone of CoA, stimulates its biosynthesis. In fact, this was observed in our studies on Ehrlich ascites tumour cells (Fig. 7). Moreover, homopantothenic acid, a derivative of pantothenic acid that is not CoA precursor, exerted a much lower protective effect (Fig. 3). The notion that pantothenic acid prevents ROS-induced apoptosis due to its action on cell glutathione is corroborated by the observations that N-acetylcysteine, a precursor of glutathione, had a similar effect (Figs. 8-11). However, the very mechanism by which increased level of CoA increases the level of cell glutathione, and possibly also its reduction state, in particular whether this proceeds by an increased biosynthesis or decreased degradation or efflux from the cell, still remains to be elucidated. From the present state of our knowledge we can only make a practical conclusion that pantothenic acid and its reduced derivative pantothenol may be important factors in prevention of oxidative stress in animal and human cells and tissues and justifies its application in medical practice and cosmetics industry.

Acknowledgements This research was supported in part by the Polish State Committee for Scientific Research under grant No. 6P04A00516 and the Jozef Mianowski Fund (Warsaw).

References [I] C. Artom, Effects of pantothenic acid and its analogs in radiation injury by *P, Proc. Expt. Biol. Med. 86 (1954) 162-165. [2] I. Szorady, Gy. Toth and I. Gazdag, X-Ray protection by means of pantothenic acid, Progr. Biochem. Pharmacol. 1 (1965) 533-536. [3] G.A. Dombradi, E. Szepesvari and I. Szorady, The possible mode of action of pantothenic acid in radioprotection, Acta Physiol. Lat. Am. 14 (1964) 327-329. [4] I. Szorady, I. Gazdag and S. Kocsor, Weitere Beobachtungen uber die Strahlenschutzwirkung der Pantothensaure, Naturwiss. 53 (1966) 527. [5] S. Casadio, A.Mantegani, G. Coppi and G. Pala, On the healing properties of esters of D-panthenol with terpene acids, with particular reference to D-pantothenyl trifamesylacetate, Arzneimittelforschung 17 (1967) 1122-1125. [6] J.F. Grenier, M. Aprahamian, C. Genot and A. Dentinger, Pantothenic acid (vitamin Bs) efficiency on wound healing, Acta Vitaminol. Enzymol. 4 (1982) 81-85. [7] M. Aprahamian, A. Dentinger, C. Stock-Damge, J.C. Kouassi and J.F. Grenier, Effects of supplemental pantothenic acid on wound healing: experimental study in rabbit. Am. J. Clin. Nutr. 41 (1985) 578-589. [8] B. Lacroix, E. Didier and J.F. Grenier, Role of pantothenic and ascorbic acid in wound healing processes: in vitro study on fibroblasts, Int. J. Vitam. Nutr. Res. 58 (1988) 407-413. [9] F. Vaxman, S. Olender, A. Lambert, G. Nisand, M. Aprahamian, J.F. Bruch, E. Didier, P. Volkmar and J.F. Grenier, Effect of pantothenic acid and ascorbic acid supplementation on human skin wound healing process. A double-blind, prospective and randomized trial, Eur. Surg. Res. 27 (1995) 158-166. [10] B.I. Weimann and D. Hermann, Studies on wound healing: effects of calcium D-pantothenate on the migration, proliferation and protein synthesis of human dermal fibroblasts in culture, Int. J. Vitam. Nutr. Res.

69(1999)113-119. [II] W. Hosemann, M.E. Wigand, U. Gode, F. Langer and I. Dunker, Normal wound healing of the paranasal sinuses: clinical and experimental investigations, Eur. Arch. Otorhinolaryngol. 248 (1991) 390-394. [12] S.F. Egger, V. Huber-Spitzy, E. Alzner, C. Scholda and V.P. Vescei, Comeal wound healing after superficial foreign body injury: vitamin A and dexpanthenol versus a calf blood extract. A randomized double-blind study, Ophthalmologica 213 (1999) 246-249. [13] E.A. Egorov, N.I. Kalinin and A.P. Kiiasov, New stimulants of comeal reparative regeneration, Vestn. Oftalmol. 115 (1999) 13-15 (in Russian with English summary). [14] T. Yoshikawa, Y. Furukawa, M. Tamai, M. Murakami and M. Kondo, The increase of lipid peroxidation in experimental hepatitis in rats induced by carbon tetrachloride or D-galactosamine and its inhibition by pantethine, J. Appl. Biochem. 4 (1982) 228-233. [15] I. Nagiel-Ostaszewski and C.A. Lau-Cam, Protection by pantethine, pantothenic acid and cystamine against

L Wojtczak and V.S. Slyshenkov I Protection by Pantothenic Acid

73

carbon tetrachloride-induced hepatotoxicity in the rat, Res. Commun. Chem. Pathol. Pharmacol. 67 (1990) 289-292. [16] L.Y. Utno, Effects of pantethine on metabolism in myocardial mitochondria under the conditions of deep hypothermia, Biull. Eksp. Biol. Med. 111 (1991) 577-578 (in Russian with English summary). [17] A.O. Kumerova, A.A. Silova and L.Y. Utno, Effect of pantethine on post-heparin lipotytic activity and lipid peroxidation in the myocardium, Biull. Eksp. Biol. Med. 111 (1991) 33-35 (in Russian with English summary). [18] A.O. Kumerova, L.Y. Utno, Z.E. Lipsberga and I.Y. Shekestere, Myocardial protection by derivatives of pantothenic acid in heart model with experimental ischemia and reperfusion, Biull. Eksp. Biol. Med. 113 (1992) 373-375 (in Russian with English summary). [19] V.M. Borets, V.A. Ovchinnikov, V.V. Mironchik, A.G. Moiseenok and M.A. Us, Pantothenic acid metabolic disorder and its relation to the change in energy processes in patients with ischemic heart disease and hypertension, Vopr. Pitan. No. 1 (1983^ 45-49 (in Russian with English summary). [20] V.M. Borets, M.A. Lis, V.M. Pyrochkin, V.P. Kishkovich and N.D. Butkevich, Therapeutic efficacy of pantothenic acid preparations in ischemic heart disease patients, Vopr. Pitan. No. 2 (1987) 15-17 (in Russian with English summary). [21] V.S. Slyshenkov, M. Rakowska, A.G. Moiseenok and L. Wojtczak, Pantothenic acid and its derivatives protect Ehrlich ascites tumor cells against lipid peroxidation, Free Radic. Biol. Med. 19 (1995) 767-772. [22] V.S. Slyshenkov, A.G. Moiseenok and L. Wojtczak, Noxious effects of oxygen reactive species on energy coupling processes in Ehrlich ascites tumor mitochondria and the protection by pantothenic acid, Free Radic. Biol. Med. 20 (1996) 793-800. [23] G. Kroemer, P.Petit P, N. Zamzami, J.L. Vayssiere and B. Mignotte, The biochemistry of programmed cell death, FASEB J. 9 (1995) 1277-1287. [24] G. Kroemer, Mitochondrial control of apoptosis: an overview, Biochem. Soc. Symp. 66 (1999) 1-15. [25] M.O. Hengartner, The biochemistry of apoptosis, Nature 407 (2000) 770-776. [26] V.S. Slyshenkov, K. Piwocka, E. Sikora and L. Wojtczak, Pantothenic acid protects Jurkat cells against ultraviolet light-induced apoptosis, Free Radic. Biol. Med. 30 (2001) 1303-1310. [27] J.L. Farber, Mechanisms of cell injury by activated oxygen species, Envir. Health Persp. 102 Suppl. 10 (1994) 17-24. [28] M. Martmez-Cayuela, Oxygen free radicals and human disease, Biochimie 77 (1995) 147-161. [29] V.S. Slyshenkov, S.N. Omelyanchik, A.G. Moiseenok, R.V. Trebukhina and L. Wojtczak, Pantothenol protects rats against some deleterious effects of gamma radiation, Free Radic. Biol. Med. 24 (1998) 894-899. [30] V.S. Slyshenkov, S.N. Omelyanchik, A.G. Moiseenok, N.E. Petushok and L. Wojtczak, Protection by pantothenol and [3-carotene against liver damage produced by low-dose y-radiation, Acta Biochim. Pol. 46 (1999)239-248. [31] S. Desagher and J.C. Martinou, Mitochondria as the central control point of apoptosis, Trends Cell Biol. 10 (2000) 369-377. [32] A. Szewczyk and L. Wojtczak, Mitochondria as pharmacological target, Pharmacol. Rev. 54 (2002) 101-127. [33] M. Zoratti and I. Szabo, The mitochondria! permeability transition. Biochim Biophys Acta 1241 (1995) 139176. [34] P. Bernard!, R. Colonna, P. Costantini, O. Eriksson, E. Fontaine, F. Ichas, S. Massari, A. Nicolli, V. Petronilli and L. Scorrano, The mitochondrial permeability transition, BioFactors 8 (1998) 273-281. [35] G. Beutner, A. Ruck, B. Riede and D. Brdiczka, Complexes between porin, hexokinase, mitochondrial creatine kinase and adenylate translocator display properties of the permeability transition pore. Implication for regulation of permeability transition by the kinases, Biochim. Biophys. Acta 1368 (1998) 7-18. [36] V. Petronilli, P. Costantini, L. Scorrano, R. Colonna, S. Passamonti and P. Bernard!, The voltage sensor of the mitochondrial permeability transition pore is tuned by the oxidation-reduction state of vicinal thiols Increase of the gating potential by oxidants and its reversal by reducing agents, J. Biol. Chem. 269 (1994) 16638-16642. [37] B.V. Chernyak and P. Bernard!, The mitochondrial permeability transition pore is modulated by oxidative agents through both pyridine nucleotides and glutathione at two separate sites, Eur. J. Biochem. 238 (1996) 623-630. [38] T. Jabs, Reactive oxygen intermediates as mediators of programmed cell death in plants and animals, Biochem. Pharmacol. 57 (1999) 231-245. [39] J. Chandra, A. Samli and S. Orrenius, Triggering and modulation of apoptosis by oxidative stress, Free Radic. Biol. Med. 29 (2000) 323-333. [40] J.M. Mates and F.M. Sanchez-Jimenez, Role of reactive oxygen species in apoptosis, Int. J. Biochem. Cell Biol. 32 (2000) 157-170. [41] A.G. Hall, The role of glutathione in the regulation of apoptosis, Eur. J. Clin. Invest. 29 (1999) 238-245. [42] M.M. Rimpler, U. Rauen, T. Schmidt, T. Moroy and H. de Groot, Protection against hydrogen peroxide cytotoxicity in rat-1 fibroblasts provided by the oncoprotein Bcl-2: maintenance of calcium homeostasis is secondary to the effect of Bcl-2 on cellular glutathione, Biochem. J. 340 (1999) 291-297. [43] V. Umansky, M. Rocha, R. Breitkreutz, S. Hehner, N. Erbe, W. Droge and A. Ushmorov, Glutathione is a factor of resistance of Jurkat leukemia cells to nitric oxide-mediated apoptosis, J. Cell Biochem. 78 (2000)

74

L Wojtczak and V.S. Slyshenkov I Protection by Pantothenic Acid

578-587. [44] J. Sastre, F.V. Pallardo and J. Vina, Mitochondria! oxidative stress plays a key role in aging and apoptosis, IUBMB Life 49 (2000) 427-435. [45] J.B. Schulz, J. Lindenau, J. Seyfried and J. Dichgans, Glutathione, oxidative stress and neurodegradation, Eur. J. Biochem. 267 (2000) 4904-4911. [46] W. Davis jr., Z. Ronai and K.D. Tew, Cellular thiols and reactive oxygen species in drug-induced apoptosis, J. Pharmacol. Exp. Ther. 296 (2001) 1-6. [47] M.R. Wieckowski, M. Vyssokikh, D. Dymkowska, B. Antonsson, D. Brdiczka and L. Wojtczak, Oligomeric Cterminal truncated Bax preferentially releases cytochrome c but not adenylate kinase from mitochondria, outer membrane vesicles and proteoliposomes, FEBS Lett. 505 (2001) 453-459. [48] M. Lutter, G.A. Perkins and X. Wang, The pro-apoptotic Bcl-2 family member tBid localizes to mitochondria! contact sites, BMC Cell Biol. 2 (2001) No. 22. [49] M. Lutter, M. Fang, X. Luo, M. Nishijima, X.-s. Xie and X Wang, Cardiolipin provides specificity for targeting of tBid to mitochondria, Nat. Cell Biol. 2 (2000) 754-756.

Thiol Metabolism and Redox Regulation of Cellular Functions A. Pompella et al. (Eds.) IOS Press, 2002

75

Thiols as Major Determinants of the Total Antioxidant Capacity A. BALCERCZYK1, A. GRZELAK1, A. JANASZEWSKA1, W. JAKUBOWSKI2, S. KOZIOL3, M. MARSZALEK1, B. RYCHLIK1, M. SOSZYNSKI1, T. BILINSKI3 and G. BARTOSZ1'3 Department of Molecular Biophysics, University of Lodz, Banacha 12/16, 90-237Lodz, Poland, fax +48 42 6354473, tel. +48 42 6354476, e-mail: gbartosz(a)Mol.uni.lodz.pl 2 Department of Biophysics, Institute of Material Engineering, Technical Univ. of Lodz; 3 Department of Cell Biochemistry and Biology, University ofRzeszow, Poland

Introduction The proper functioning of aerobic organisms subjected to the danger of uncontrolled oxidation depends critically on the efficient antioxidant system. This system includes antioxidant enzymes, low-molecular weight antioxidants and biological chelators of transition metal ions. What is an antioxidant? According to the apparently most useful definition, antioxidant is "any substance that, when present at low concentrations compared with those of an oxidizable substrate, significantly delays or prevents oxidation of that substrate" [1]. The usefulness of this definition lies, i. a., in its instrumentality as it suggests straightforward means of testing antioxidant properties of chemical compounds. In fact, testing the delay or inhibition of a chosen oxidation reaction allows for a simple measurement of the sum of all antioxidants present in the sample studied. This parameter is usually referred to as Total Antioxidant Activity or, as recommended lately, Total Antioxidant Capacity (TAC) [2]. Total Antioxidant Capacity is an increasingly frequently measured parameter. While lacking detailed information about the composition and concentrations of individual antioxidant species present in a sample, it provides an easy and rapid means of evaluation of "antioxidant power" of the examined material. An obvious field of application of TAC assay is analysis of plants, food products and beverages. However, there are many attempts to apply this analysis for evaluation of antioxidant status of body fluids (especially blood plasma) in normal human subjects and animals. For research purposes, it is also of interest to evaluate TAC of cells and tissue homogenates in animals subjected to different treatments. There are numerous methods of estimating TAC. One group of methods is based on the inhibition of free-radical induced oxidation of an indicator compound or of a substrate yielding an indicator [3-5]. The indicator is usually chosen to have characteristic color or fluorescence. Another group of assaying TAC is based on the property of most antioxidants to reduce stable free radicals [6, 7] or ferric ions (FRAP, Ferric Reducing Activity of Plasma) [8, 9]. These reductive assays are simple, less time-consuming and less troublesome in reading the results so their popularity is increasing. In this study we employed mainly the method of reduction of preformed 2,2'-azinobis(3ethylbenzthiazoline-6-sulfonic acid) (ABTS) cation radical (ABTS*+) [7] and FRAP [8, 9]. Various methods of TAC assay employ different oxidants (if any), different indicators and different reaction conditions. It is not surprising, therefore, that the results

76

A. Balcerczyk et al. / Thiols as Determinants of the Antioxidant Capacity

obtained with various assays differ to some extent as do reported contributions of individual antioxidants to TAG. The material studied most extensively for TAG is blood plasma. The main contributor to TAG of human blood plasma is uric acid (Table 1). Table 1. Contribution of main antioxidants to the Total Antioxidant Capacity of blood plasma [%]. ABAP, 2,2'-azobis(2-amidopropane) hydrochlorkte. Urate

Protein thiots

VttamhiE

Ascorb*.

Reference

58 ±18

21 ±10

7±2

14 ±8

[10]

ABAP/RPhycoerythrin

36.9 ± 13.1

31 .5 ±13.1

4.8 ±1.8

3.7 ±0.9

111]

ABAP/Luminol

43.6 ±6.9

18.3 ±3.6

4.4 ±1.5

2.1 ±1.2

112]

50.9

20.8

1.6

26.7

[13]

60

10

5

15

[91

Method TRAP

Cretin bleaching FRAP

Fig. 1. Effect of modification of thiol content on the TAG of bovine serum albumin (BSA). BSA (50 mg/ml in phosphate-buffered saline, pH 7.4) was treated with diamide (1 mM), Nethylmaleimide (NEW; 1 mM) or dithiothreitol (DTT; 1 mM) for 1 h at room temperature. Excess of the reagents was removed by overnight dialysis. Thiol groups were estimated with the EHman reagent [17]. TAG was estimated by slightly modified methods of ABTS** bleaching [71 and TRAP [8]. Our modification of the method of ABTS** bleaching consists in the use of ABTS* solution of A414 = 1 in 10 mM sodium phosphate. pH 7.4, and measurement of absorbance at 414 nm usually after 10 s (here also after 1 minute). We measured TRAP after 20-min reaction at ambient temperature.

A. Balcerczyk et al. I Thiols ax Determinants of the Antioxidant Capacity

77

120 T

2

8

100

40

Q

20

S

5

3 - 3 .

Fig. 2. Effect of experimental modification of the thtol content on TAG of Mood plasma. Thiols were estimated according to Ellman [17]. TAG was estimated by ABTS** bleaching after 10s.

However, protein thiols are the second biggest component contributing to TAG estimated by most methods (FRAP being an exception, due to the low reactivity of thiols with ferric ions). The antioxidant power of uric acid is well documented; however, the sequence of

78

A. Balcerczyk et al. / Thiols as Determinants of the Antioxidant Capacity

consumption of blood plasma antioxidants is: ascorbate = protein thiols > bilirubin > uric acid > alpha-tocopherol. It means that protein thiols are much more sensitive to oxidative stress than uric acid and changes in TAG should correlate with alterations in the level of thiols under conditions of moderate oxidative stress. Such a situation has indeed been observed in many cases including, i. a., fetal hypotrophy (Karowicz et al., submitted). On the other hand, changes in the level of uric acid were sometimes poorly correlated with oxidative stress; especially, increased TAG was observed in critically ill patients patients with renal dysfunction, due to increase in uric acid level [14]. Attempts have been made to calculate and use the urate-independent fraction of TAG as a better predictive of whole-body antioxidant status [15]. Thiols of the blood plasma belong mostly to plasma proteins. Bovine serum albumin as a model plasma protein shows TAG values comparable with the thiol content. However, even in this simple system, TAG values obtained from measurements of ABTS*+ reduction depend heavily on the time of the measurement. ABTS*+.is known to react rapidly with thiols and to show a sluggish but promiscuous activity with other amino acid residues, especially those of tyrosine and tryptophan [16]. Although the TAG values measured by ABTS*+ reduction after 1 min were most closely related to the thiol content of the protein, those obtained by measurements after 10 s were found to correlate better with the thiol content (Fig. 1). Experimental modification of the thiol level of blood plasma had little effect on TAG of the plasma (Fig. 2). Estimation of TAG of tissue and cell homogenates is an interesting field of research. TAG of such material can be expected to be in most cases even more dependent on the contribution of -SH groups than that of blood plasma due to the high intracellular concentration of glutathione and protein thiols. Our results showed that in hemolysates of human erythrocytes TAG measured by ABTS*+ reduction after 10s was about two times higher than the thiol group content and was decreased by about 1/3 when -SH groups were practically totally oxidized with diamide or blocked with NEM. The moderate dependence of TAG of hemolysates on the thiol content is due to the reactions of hemoglobin since a much stronger dependence was noted after separation of hemoglobin with Centricone (Fig. 3)TAG of cell extracts of the budding yeast Saccharomyces cerevisiae showed a stronger dependence on the thiol content as judging from the effect of -SH blocking with NEM (Fig. 4). In this case, TAG measured after 10 s was decreased by 83-90% (in different strains) after thiol modification while TAG measured after 1-min reduction of ABTS*+ was decreased by 73-80%. These results indicate that thiol groups are a major contributor to TAG of yeast cells. Comparison of the thiol content and TAG of several lines of cultured mammalian cells lead to a similar conclusion. In this case, since the amount of material was limited, we did not use the Ellman method of thiol determination in cell extracts but, instead, measured thiol groups in intact cells by electron spin resonance (ESR) using a bis-(2,2,5,5)tetramethyl-3-imidazoline-l-oxyl-4-yl) disulfide biradical (RSSR) spin label. This spin label is a disulfide and its reaction with thiols reduce intramolecular spin-spin interactions and increase in the ESR signal [18] (Fig. 5). Depletion of cellular thiol groups by treatment with diamide or NEM decreased TAG of cell extracts (measured by ABTS*+ reduction) by up to 89% (Fig. 6).

A. Balcerczyk et al. / Thiols as Determinants of the Antioxidant Capacity

79

Fig. 3. Effect of modification of thiol content on the TAG of hemotysates (HEM) and hemoglobin-free Centricone filtrate* of hemolysates (FIL) of human erythrocytes. Erythrocytes (hematocrit of 0.10 in PBS) were treated with diamide (1 mM) or N-ethylmaleimide (NEM; 2 mM) for 1 h at room temperature. Thiol groups were estimated with the Ellman reagent [17]. TAG was estimated by ABTS** reduction [7] after 10 s. Absorbance was measured at 734 nm to avoid interference with light absorption of hemoglobin at shorter wavelengths.

These results demonstrate that in cell extracts, in contrast to extracellular fluids, thiol groups constitute the dominant determinant of Total Antioxidant Capacity. Depletion of thiols leads to decrease of TAG. However, cellular adaptation to oxidative stress may involve mobilization of other mechanisms than increase of thiol concentration. This is especially evident in yeast cells where strains deficient in antioxidant enzymes show increased values of TAG due mainly to thiol-independent mechanisms [19, 20]. Similarly, adaptation of yeast to conditions of stationary culture (Fig. 7) and reoxygenation after growth in anoxia (Fig. 8) involve mainly other antioxidants than thiols. These data are in line with those of Evelson et al. on TAG of tissue homogenates showing that glutathione is the main factor affecting TAG and that about half of TAG of tissue homogenates is not accounted for and is due to reactions of proteins [21]. Therefore, within the cells low-molecular weight thiols and protein thiol groups constitute a major but not always decisive determinant of TAG.

80

A. Balcerczyk et at. / Thiols as Determinants of the Antioxidant Capacity

Fig. 4. Effect of NEM (2 mM) on the content of total and acid-soluble thtols and TAG estimated by ABTS** reduction after 10 s. Cells from logarithmic cultures in yeast extractpeptone-glucose medium (control or treated with 2 mM NEM for 30 min) were broken with glass beads. Acid-soluble fraction was obtained by treatment with trichloroacetic acid (5% final) and analysed after neutralization in phosphate buffer. Thiol groups were estimated with Eflman reagent and TAG by reduction of ABTS**. AS -SH, acid-soluble thiols; T -SH, total thiols. NEM - N-ethylmaleimide-modified cells. SP4. wild-type strain; SOD1-, strain devoid of CuZnSOD; SOD1-SOD2-, strain devoid of CuZnSOD and MnSOD; CATA-CATT-, strain devoid of catalase A and catalase T.

A. Balcerczyk et al. / Thiols as Determinants of the Antioxidant Capacity

Fig. 5. ESR spectra of the RSSR spin label (100 uM) in PBS (left) and after reaction with 10 uM glutathione. The height of the main triplet of the spectrum increases after reaction with thiols. RSSR was purchased from Alexis (Switzerland).

Fig. 6. Effect of diamide and NEM on the thiol group content (a) and TAG (b) of various cell lines. Thiol content was measured with RSSR spin label in whole cells and TAG was estimated by reduction of ABTS*+in cell extracts obtained by lysis of the cells with 100 uM digitonin. TE, Trolox equivalents.

82

A. Balcerczyk et al. / Thiols as Determinants of the Antioxidant Capacity

Fig. 7. Changes in the concentration of acid-soluble thtols (AS -SH) and TAG of S. cerevis/ae cells grown on yeast extraxt-peptone-galactose medium during transition from logarithmic (log) to stationary (sta) culture. TAG was estimated by FRAP and reduction of 1,1'-diphenylpicrylhydrazyl (DPPH) free radical [6].

A. Balcerczvk et a!. / Thiols as Determinants of the Antioxidant Capacity

83

Fig. 8. Changes in the concentration of acid-soluble thiols (AS -SH) and TAG of S. cerev/siae cells grown on yeast extract-peptone-glucose medium after oxygenation of anaerobic logarithmic cultures.

References [1] B. Halliwell and J. M. C. Gutteridge, Free Radicals in Biology and Medicine. Oxford University Press, Oxford, New York, 3rd Edition, 1999. [2] A. Ghiselli, M. Serafmi, F. Natella and C. Scaccini, Total antioxidant capacity as a tool to assess redox status: critical view and experimental data, Free Radic. Biol. Med. 29 (2000) 1106-1114. [3] D. D. M. Wayner, G. W. Burton, K. U. Ingold and S. Locke, Quantitative measurement of the total peroxyl radical trapping antioxidant capbility of human blood plasma by controlled peroxidation. The important contribution made by plasma proteins, FEBS Lett. 187 (1985) 33-37. [4] G. Cao, H. M. Alessioo and R. G. Cutler, Oxygen-radical absorbance capacity assay for antioxidants, Free Radic. Biol. Med. 14(1993)303-311. [5] C. Rice-Evans and N. J. Miller, Total antioxidant status in plasma and body fluids, Meth. Enzymol. 234 (1994)279-293. [6] M. S. Blois, Antioxidant determinations by the use of a stable free radical, Nature 181 (1958) 1199-1200. [7] R. Re, N. Pellegrini, A. Proteggente, A. Pannala, M. Yang and C. Rice-Evans, Antioxidant activity applying an improved ABTS radical cation decolorization assay, Free Radic. Biol. Med. 26 (1999) 1231-1237.

84

A. Balcerczyk et al. / Thiols as Determinants of the Antioxidant Capacity

[8] I. F. F. Benzie and J. J. Strain, The ferric reducing ability of plasma (FRAP) as a measure of "antioxidant power": the FRAP assay, Anal. Biochem. 239 (1996) 70-76. [9] I. F. Benzie and J. J. Strain, Ferric reducing/antioxidant power assay: direct measure of total antioxidant activity of biological fluids and modified version for simultaneous measurement of total antioxidant power and ascorbic acid concentration, Methods Enzymol. 299 (1999) 15-27. [10] D. D. Wayner, G. W. Burton, K. U. Ingdd, L R. Barclay and S. J. Locke, The relative contributions of vitamin E, urate, ascorbate and proteins to the total peroxyl radical-trapping antioxidant activity of human blood plasma. Biochim. Biophys. Acta 924 (1987) 408-419. [11] A. Ghiselli, M. Serafini, G Maiani, E. Azzini and A. Ferro-Luzzi, A fluorescence-based method for measuring total plasma antioxidant capability, Free Radical Bid. Med. 18 (1995) 29-36. [12] M. Ertiola, M. M. Nieminen, P. Kellokumpu-Lehtinen, T. Metsa-Ketela, T. Poussa and H. Alho. Plasma peroxyl radical trapping capacity in lung cancer patients: a case-control study, Free Radic. Res. 26 (1997) 439-447. [13] F. Tubaro, A. Ghiselli, P. Rapuzzi, M. Maiorino and F. Ursini, Analysis of plasma antioxidant capacity by competition kinetics, Free Radic. Biol. Med. 24 (1998) 1228-1234. [14] K. L MacKinnon, Z. Molnar, 0. Lowe, I. D. Watson and E. Shearer, Measures of total free radical activity in critically ill patients, Clin. Biochem. 32 (1999) 263-268. [15] I. N. Popov and G. Lewin, Antioxidative homeostasis: Characterization by means of a chemiluminescent technique, Meth. Enzymol. (1999)437-456. [16] G. Bartosz and M. Bartosz, Antioxidant activity: what do we measure?, Acta Biochim. Pol. 46 (1999) 23-

29. [17] G. Ellman, Tissue sulfhydryl groups, Arch. Biochem. Biophys. 82 (1959) 70-77. [18] L M. Weiner, Quantitative determination of thiol groups in low and high molecular weight compounds by electron paramagnetic resonance, Meth. Enzymol. 251 (1995) 87-105. [19] E. Jaruga, E. A. Lapshina, T. Bilihski, A. Ptonka and G. Bartosz, Resistance to ionizing radiation and antioxidative defence in yeasts. Are antioxidant-deficient cells permanently stressed?, Biochem. Mol. Bid Int. 37(1995)467-472. [20] E. A. Lapshina, E. Jaruga, T. Bilihski and B. G., What determines the antioxidant potential of yeast cells?, Biochem. Mol. Bid. Int 37(1995) 903-908. [21] P. Evelson, M. Travacio, M. Repetto, J. Escobar, S. Uesuy and E. A. LJssi, Evaluation of total reactive antioxidant potential (TRAP) of tissue homogenates and their cytosols, Arch Biochem Biophys 388 (2001)

261-266.

Thiol Metabolism and Redox Regulation of Cellular Functions A. Pompella et at. (Eds.) IOS Press, 2002

85

Enzymes of the Thiol-Dependent Hydroperoxide Metabolism in Pathogens as Potential Drug Targets Heike BUDDE and Leopold FLOHE Dept. of Biochemistry, Technical University of Braunschweig Mascheroder Weg 1, D-38124 Braunschweig, Germany

1. Introduction Infectious diseases are still the major threat to human health and are estimated to cause 13 million fatalities per year. The need of new antibiotics does not only result from the developement of resistance against available drugs. The possibilities to treat parasitic diseases prevailing in poor people of poor countries have for ever been unsatisfactory and are evidently not improving. Obviously, the substantial investments required to develop new therapies according to present standards do not seem economically rewarding [1]. There is no shortage of science-based ideas how the situation could be changed. The analyses of whole genomes and proteoms of pathogens disclose a realm of structures and metabolic pathways that are unique to pathogens, and their relevance to pathogenicity and virulence can now be investigated, e. g., by inverse genetics with a reliability and speed that could not be dreamed of a few decades ago [2]. The present article tries to work out the options only arrising from a single aspect of host/pathogen interaction, which, however, is relevant to infectious processes in general.

2. Detoxification of hydroperoxides in mammals and pathogens 2.1. The Needs Whenever a microorganism intrudes into mammalian tissue, recognition of a pathogen by phagocytes triggers the release of -Oa" by activation of NADPH oxidase [3,4]. The superoxide anion, at the low pH of the phagosome, is spontaneously dismutated to molecular Oa and F^O; at neutral pH dismutation of -C>2~ is achieved by superoxide dismutases. Also, -OjT reacts with phagocyte-derived -NO to yield peroxynitrite [5]. Myeloperoxidase uses HiOa to produce an even stronger oxidant, hydrochlorite [6]. Further, FJbOi in the presence of transition metals will generate the most aggressive -OH. Simultaneously, hydroperoxides of unsaturated fatty acids will be formed by lipoxygenases or free-radical chains sustained by FfcOa and other hydroperoxides. Like NADPH oxidase the lipoxygenases remain dormant as long as not activated by complex signalling cascades. They further require hydroperoxides to become active. Their ultimate products, with hydroperoxy and hydroxy fatty acids, prostaglandins and leukotrienes, comprise the lipid mediators known to trigger and amplify inflammatory responses in concerted actions with cytokines [4,7].

86

H. Budde and L. Flohe / Enzymes in Pathogens as Potential Drug Targets

These common responses to almost everything recognized as foreign are ment to eliminated intruded pathogens but have also to be seen as friendly fire damaging the host's own cells. In extreme cases, like septicemia, where microorganisms are spread all over the organism, the host organism may die, not necessarily from toxins of the pathogens, but from an exaggerated response of its own defense system. Since H2C>2 and other hydroperoxides, being oxidants themselves, as well as signals to produce inflammatory mediators and sources of damaging free radical chains, are central components of the host defense reaction, both, the affected host and the pathogen had do develop strategies for their elimination. As is to be demonstrated, these strategies vary appreciably in phylogenetically remote domains. This offers the opportunity of selective inhibition of the pathogen's defense mechanism which would limit its ability to survive in the hostile environment, while leaving the host's self-protection unimpaired. [4,8-1 1] 2.2. Self-Protection of the Mammalian Host The best known HiCh eliminating enzyme is catalase which dismutates H2C>2 to H2O and molecular oxygen [12]. Its relevance in the context of the host defense reactions is questionable. In all highly structured mammalian cells catalase appears to be restricted to peroxisomes, where it is involved in detoxifying FhCh derived from peroxisomal metabolism [13]. At least equally efficient in H2O2 elimination are the selenium-containing glutathione peroxidases (GPx). They also have the advantage to reduce a wide range of organic hydroperoxides [14]. Amongst these selenoperoxidases, the ubiquitous GPx-1 present in the cytosol and the mitochondria! matrix appears to be most relevant to general antioxidant defense, as is required in an oxidative stress resulting from an infection. Treatment with bacterial lipopolysaccharides, an accepted model to mimick septicemia, killed GPx (-/-) mice more readily than wild-type mice [15]. GPx-4, the phospholipid hydroperoxide glutathione peroxidase, was shown to silence 5-lipoxygenase thereby preventing leukotriene formation and consecutive inflammatory responses [16]. It further dampens the response to inflammatory cytokines such as IL-1 [17]. GPx-3, the extracellular form, may be discussed to regulate inflammatory responses by lowering the peroxide tone in the extracellular space. It there is operating at high efficiency but low capacity, because the supply of reducing equivalents, which are GSH or thioredoxin, is limited. It thus is in an ideal condition to dampen an irrelevant inflammatory stimulus, while, upon consumption of thiol substrates, it allows a full host defense reaction if the stimulus is strong enough, i. e. demanding a serious response. This appealing hypothesis [14], however, awaits experimental verification. The role of GPx-2, largely restricted to the epithelial lining of the gastro-intestinal system, still remains elusive. However, mice deficient in both GPx-1 and GPx-2 spontaneously develop symptoms reminding of Crohn's disease [18] suggesting a self-protecting role of the enzymes in the steady, usually successful fight against the intestinal flora. The selenoperoxidases were also shown to reduce peroxynitrite in vitro [19]. But surprisingly, GPx-1 (-/-) mice tolerated a peroxynitrite challenge better than wildtype mice [20]. Their ability to reduce peroxynitrite is shared by low molecular weight selenols, seleno-ethers and other selenoproteins [21]. Selenoprotein P with its up to 12 selenocysteine residues is now discussed as the more efficient "peroxynitrite reductase". Its extracellular localisation in conjunction with its affinity to sulfatated glycans at the endothelial surfaces further support to the idea that is may protect the endothelium against macrophage-derived peroxynitrite [22]. More recently, also mammalian peroxiredoxins have attracted considerable attention as peroxide detoxifying systems. Depending on the particular type, they can reduce hydroperoxides at the expense of thioredoxins [23] or GSH [24]. As discussed in detail elsewhere, their molar efficiency can not likely compete with that of the selenoenzymes,

H. Budde and L Flohe / Enzymes in Pathogens as Potential Drug Targets

because the sulfiir catalysis, they depend on, is substantially slower than selenium catalysis [25]. More likely, therefore, the specific roles of mammalian peroxiredoxins are to be sought in the redox regulation of particular cellular events such as signalling and differentiation.

Fig.1. Thlol-mediated hydroperoxide metabolism in mammals. Homologous proteins are shown in identically marked circles. Se indicates redox-active selenium. TrxR, thioredoxinreductase; GR, glutathione reductase; Trx, thioredoxin; Grx, glutaredoxin; Prx, peroxiredoxin; GPx, glutathione peroxidase.

2.3. Hydroperoxide Metabolism in Trypanosomatids Pathogenic trypanosomatids like Trypanosoma and Leishmania species do neither contain catalase nor the selenium-type glutathione peroxidases [26]. They can synthetize GSH but transform most of it to a bis-glutathionyl derivative of spermidine called trypanothione [T(SH)2]. The synthesis of T(SH)2 is achieved by the two distinct, though related, enzymes: glutathionyl-spermidine synthetase (GSS) and trypanothione synthetase (TS). Which of the published sequences (TrEMBL AC: P90518, O60993, Q9GT49, Q9GT48) is correctly attributed to which of the two enzymatic activities is still debated [27,28]. Unlike mammals, trypanosomatids are unable to reduce oxidized GSH (GSSG) at the expense of NADPH. GSSG reduction may be achieved, however, by T(SH)a, whereby oxidized trypanothione TS2, which is an energetically preferred cyclic disulfide, is formed. The reaction is likely catalysed by a transhydrogenase [29]. T(SH)2 then is regenerated from TS2 with NADPH by trypanothione reductase (TR), a flavine-containing enzyme that is homologous to glutathione reductase (GR). T(SH)2 is the main redox mediator to detoxify hydroperoxides in trypanosomatids. This reaction had for long been assumed to be homologous or at least analogous to the GPx reaction. The hypothetical "trypanothione

88

H. Budde and L. Flohe / Enzymes in Pathogens as Potential Drug Targets

peroxidase", however, could never be identidified. Instead, the reaction turned out to require two distinct catalysts, a thioredoxin-related protein called tryparedoxin (TXN) and a peroxiredoxin-type peroxidase called tryparedoxin peroxidase (TXNPx) [30]. The comparatively low molar efficiency of TXNPx is compensated for, in Crithidia fasciculata at least, by extremely high concentrations amounting to more than 5 % of the soluble protein [30]. This unique pathway of peroxide detoxification was first discovered in the insect-pathogen C. fasciculata [30] but is evidently common to all the pathogenic species of the kinetoplastida family, e. g. T. brucei brucei [31], T. brticei rhodesiense [32], T. cruzi [33-35], L. major [36], L. donovani [37], and L. infantum (Castro et.al, unpublished). The system was unequivocally shown to be relevant to hydroperoxide detoxification in trypanosomatids. Conditioned knock out of TR in T. brucei brucei resulted in dramatically increased FfcCh-sensitivity in vitro and loss of virulence in experimentally infected mice [38], and a dominant negative approach to reduce TR activity in L. donovani impaired survival in macrophages [39]. TR may thus be considered as a molecular target for the developement of trypanocidal drugs that has been validated according to the state-of-theart, and the downstream enzymes of the system, as well as those involved in T(SH>2 synthesis, may reasonably be presumed to be of equal vital importance to the parasite. Over the past two years, however, the seemingly established trypanosomal hydroperoxide metabolism has been enriched by further complexities, i) TXN appears to serve multiple purposes. Apart from being the reducing substrate for TXNPx, it was shown to also support ribonucleotide reductase of T. brucei, thus being relevant to DNA synthesis [40]. ii) a typical thioredoxin was discovered in T. brucei raising the problem of mutual substitution of these redox catalysts [41]. iii) While the hydroperoxidase system, as described for C. fasciculata, was clearly localized to the cytosol [42], mitochondria! isozymes of TXNPx were discovered in T. cruzi [35], T. brucei [31] and L. infantum (Castro et al., unpublished), iv) Two members of the GPx family were discovered in T. cruzi, both being cysteine homologs of the selenocysteine-containing mammalian prototypes, and v) one of them, 7cGPx-l proved to prefer TXN over GSH as reducing substrate, thus exemplifying that relatedness in sequence may be misleading in predicting enzymatic function [43]. Taken together, these puzzling findings demand further studies on the precise role of the individual proteins before they can reliably by rated as drug targets. 2.4.Antioxidant Defense in Other Protozoa! Parasites As reviewed recently [10], Plasmodium species are equipped with all enzymes to synthetise and regenerate GSH, and GSH has for long been considered to be the most relevant antioxidant mediator of these parasites. Also a GPx-type protein was discovered in P. falciparum [44]. This GPx, which again is not a selenoprotein, expectedly proved to be a moderately efficient peroxidase, but surprisingly accepted plasmodial thioredoxin as reducing substrate, while GSH, as judged from infinite Km and low rates, behaved like an unspecific foreign substrate [45]. Further, P. falciparum contains two peroxiredoxins [46] and a glutaredoxin [47]. One of the peroxiredoxins at least proved to be a typical thioredoxin peroxidase [46], while the role of the second remains to be worked out. In short, then, P. falciparum is equipped with a thioredoxin-mediated hydroperoxide detoxification system. It is fueled by a typical thioredoxin reductase (TrxR), which is a member of the FAD-containing disulfide reductases and is homologous to the mammalian TrxR, although distinct in not containing selenocysteine. The Trx provides the reduction equivalents to two peroxidases belonging to different protein families. In view of the novel insights the suspected relevance of GSH to plasmodial antioxidant defense remains an intriguing problem.

H. Rudde and L. Flohe / Enzymes in Pathogens as Potential Drug Targets

Reductant

89

NADPH

Flavoproteins LMW Mediators

CXXCproteins Substrates

Peroxidases

Products

Fig. 2. Thiol-mediated hydroperoxide metabolism In Trypanosomtlds. Homologous proteins are marked as in Fig. 1. TryR, trypanothione reductase; TXN, tryparedoxin; TXNPx, peroxiredoxin-type tryparedoxin peroxidase; GPxl and II, GPx-type peroxidases. The relevance of the thioredoxin (Trx) pathway to hydroperoxide detoxification in trypanosomatids is unknown.

Reductant

NADPH

Flavoproteins LMW Mediator CXXCproteins

Substrates

Peroxidases

Products

Fig. 3. Thiol-medlated hydroperoxide metabolism In Plasmodlum. Description of proteins corresponds to Figs. 1 and 2. Trx reduces both, Prx- and GPx-type enzymes. The metabolic context of the one-cysteine Prx (1-Cys-Prx) and the relevance of the glutathione system to antioxidant defense is unclear [47].

In contrast to Plasmodium species, Entamoeba histolytica is reportedly unable to synthetise GSH [48]. Surprisingly, however, the parasite was claimed to contain trypanothione [49] and trypanothione synthetase activity and is presumed to produce T(SH>2 by means of GSH taken up from host cells [50]. E. histolytica contains a peroxiredoxin-type peroxidase [51]

90

H. Budde and L. Flohe / Enzymes in Pathogens as Potential Drug Targets

that, qualitatively at least, accepts heterologous rat thioredoxin [52] and tryparedoxin of C. fasciculata [25], but not AhpF of Salmonella typhimurium [52]. The endogenous reductant of the peroxiredoxin in Amoeba remains to be elucidated. Organism Reductant

H. pylori

E. coli, S. typhi

M. tuberculosis

NADPH

NADPH

Flavoprotein

Mediator Level

Products

Fig. 4. Examples of bactertal antloxtdant defense systems. Proteins involved are marked for homotogy as in Fig. 1. H. pylori makes use of the Trx-system for reduction of Prx-type AhpC like mammals and many others organism. In Enterobacteria AhpC is directly reduced by a flavindependent disurfide reductase. They also contain GSH and (non-Se) GPx homotogues of unclear function (not shown). In Mycobacteria AhpC reduction appears to be linked to energy metabolism (see text). This family also contains a thioredoxin system, further glutaredoxin-like proteins and mycothiol as low molecular mass redox mediator. The role of these redox systems in antioxidant defense is still unclear.

2.5. Diversified Alkyl Hydroperoxide Reductase Systems in Bacteria Bacterial hydroperoxide reduction is known to be catalysed by peroxiredoxin-type peroxidases called AhpC. In enterobacteria AhpC, together with the flavine-containing disulfide reductase AhpF, forms a two-component system achieving hydroperoxide reduction by NAD(P)H without the aid of auxiliary enzymes or low molecular weight redox mediators [53,54]. In E. coli and S. typhimurium AhpF and AhpC are jointly regulated by the oxyR regulon that responds to oxidative stress and determines hydroperoxide resistance [55,56]. While AhpC appears to be ubiquituous in bacteria [25], AhpF is not. Helicobacter pylori lacks AhpF. Instead, H. pylori uses a thioredoxin as the reductant of AhpC [57]. The most important pathogen Mycobacterium tuberculosis contains an AhpC gene that is overexpressed and appears to be a virulence factor in catalase-negative isoniazid-resistant strains [58]. In contrast to enterobacteria, M. tuberculosis does not respond to oxidative stress with overexpression of AhpC [55]. Mycobacteria do neither contain a functional AhpF.

H. Budde and L Flohe / Enzymes in Pathogens ax Potential Drug Targets

91

Recently, a quite unique pathway of AhpC reduction in M. tuberculosis has been proposed [59]: The reduction equivalents are transferred from NADH by two components of the o> keto acid dehydrogenase complex, dihydroliponamide dehydrogenase (Lpd) and dehydroliponamide succinyltransferase (SucB), to AhpD that, like thioredoxins, contains a CXXC motif and thus substitutes for a real thioredoxin in AhpC reduction. The low molecular weight thiol that often mediates the flux from the flavoproteins to CXXC proteins is here replaced by SucB-bound lipoic acid. It remains to be worked out whether this extraordinary route, that links antioxidant defense to energy metabolism, is the only or the most relevant mechanism of peroxide detoxification in mycobacteria. They contain various thioredoxins, glutaredoxin-like proteins and the small redox mediator mycothiol (but no GSH) [60]. How these redox-active compounds are integrated in mycobacterial antioxidant defense awaits clarification. As evident from the examples mentioned, hydroperoxide metabolism in bacteria proved to be highly diversified and more surprises can be expected. Also, protein families not uncommon in mammalian hosts such as heme peroxidases and GPx-type enzymes may contribute to bacterial antioxidant defense, although the GPx-type proteins identified so far in bacteria are all cysteine homologues of the more efficient selenocysteine-containing enzymes of mammals [14].

3. Summarizing Facts and Fictions • •

•

•

• • •

• •

Clearly, hydroperoxide metabolism in pathogens often differs substantially from that in mammalian hosts. The components of the detoxification systems are composed of phylogenetically related proteins. The diversification in structures and specificities is, however, pronounced enough to justify the attempt to search for specific inhibitors. The relevance of some unique microbial pathways to virulence has been documented, e. g. for the trypanothione-mediated one in T. brucei brucei [38] and L. donovani [39] and for AhpC of isoniazid-resistant strains of M tuberculosis [58]. The idea to exploit such pathogen-specific detoxification systems as drug targets is underscored by drugs that interfer, though not selectively enough, with such pathways, e. g. difluoromethylornithine impairing spermidine synthesis and arsenicals interacting with tryparedoxins [30]. Whole genomes, being established as for P. falciparum or M. tuberculosis or being close to completetion as for trypanosomes greatly facilitate the identification of novel targets. The knowledge of relevant genes guarantees an easy production for most of the proteins required for large-scale testing. Reaction mechanisms, substrate specificities and kinetics have been elucidated for most of the targets in question [10,37]. On this basis, analytical procedures can be adapted to high throughput screening without forseeable obstacles. The structures of proteins typical of both, the microbial and the mammalian systems are emerging fast enough to pave the way for rational drug design [25,59,61-65]. The scientific and technical prerequisites thus being excellent, the chances to improve anti-infectious therapy by systematically exploiting the pathogens' idiosyncasies in antioxdant defense should no longer be missed.

92

H. Budde and L Flohe / Enzymes in Pathogens as Potential Drug Targets

Acknowledgements The work on hydroperoxide metabolism of trypanosomatids was supported by the DFG (Grants Fl 61/8-1,2, 3 and Fl 61/11-1, 2, 3).

References [1] M.L Cohen, Changing patterns of infectious disease. Nature 406 (2000) 762-767. [2] B.R. Bloom, On the particularity of pathogens, Nature 406 (2000) 760-761. [3] B.M. Babior, R.S. Kipnes, J.T. Cumutte, Biological defense mechanisms: the production by leukocytes of superoxkJe, a potential bacterial agent, J. Clin. Invest. 52 (1973) 741-744. [4] B.M. Babior, Phagocytes and oxidative stress, Am. J. Med. 109 (2000) 33-44. [5] J.S. Beckman, W.H. Koppenol, Nitric oxide, superoxide and peroxynitrite: the good, the bad and the ugly, Am. J. Phystol. 271 (1996) C1424-C1437. [6] S.J. Klebanoff, H. Rosen, The role of myetoperoxkJase in the microbicidal activity of pdymorphonudear leukocytes. In: C.F. Symp. (Ed.), Oxygen free radicals and tissue damage. Excerpta Media, Amsterdam/Oxford/NewYork, 1978, pp. 263-284. [7] R. Brigelius-Flohe, M. Maiorino. F. Ursini, L. Flohe, Selenium: An Antioxidant? In: E. Cadenas, L. Packer (Eds.), Handbook of Antioxidants. Marcel Dekker, Basel, New York, 2002, pp. 633-664. [8] L Flohe, The Achilles' heel of trypanosomatids: trypanothione-mediated hydroperoxide metabolism, Biofactors 8 (1998) 87-91. [9] A.M. Fairlamb, Future prospects for the chemotherapy of Chagas' disease, Medicina (B Aires) 59 (1999) 179-187. [10]L Flohe, H.J. Hecht, P. Steinert, Glutathione and trypanothione in parasitic hydroperoxide metabolism, Free Rad. Biol. Med. 27 (1999) 966-984. [11]R.L. Krauth-Siegel, G.H. Coombs, Enzymes of parasite thiol metabolism as drug targets, Parasftol. Today 15(1999)404-409. [12JG.R. Schonbaum, B. Chance, Catalase. In: P.O. Boyer (Ed.), The Enzymes. Academic Press, New York, San Francisco, London, 1976, pp. 363-408. [13]H. Sies, C. Gerstenecker, H. Menzel, L Flohe, Oxidation in the NADP system and release of GSSG from hemoglobin-free perfused rat liver during peroxidatic oxidation of glutathione by hydroperoxides, FEBS Lett 27(1972)171-175. [14]L Flohe, R. Brigelius-Flohe, Selenoproteins of the glutathione system. In: D.L HatfiekJ (Ed.), Selenium. Its Molecular Biology and Role in Human Health. Kluwer Academic Publishers, Boston/Dordrecht/London, 2001, pp. 157-178. [15]H. Jaeschke, Y.-S. Ho, M.A. Fischer, J.A. Lawson, A. Farhood, Glutathione peroxkjase-deficient mice are more susceptible to neutrophil-mediated hepatic parenchymal cell injury during endotoxemia: importance of an intracellular oxidant stress, Hepatotogy 29 (1999) 443-450. [16]F. Weitzel, A. Wendel, Setenoenzymes regulate the activity of leukocyte 5-lipoxygenase via the peroxide tone, J. Biol. Chem. 268 (1993) 6288-6292. [17]R. Brigelius-Ftohe, B. Friedrichs, S. Maurer, M. Schultz, R. Stretcher, lnterieukin-1-induced nuclear factor kappa B activation is inhibited by overexpresskm of phospholipid hydroperoxide glutathione peroxidase in a human endothelial cell line, Biochem. J. 328 (1997) 199-203. [18JR.S. Esworthy, R. Aranda, M.G. Martin, J.H. Doroshow, S.W. Binder, F.F. Chu, Mice with combined disruption of Gpx1 and Gpx2 genes have colitis, Am. J. Physiol. Gastrointest. Liver Physio). 281 (2001) G848-G855. [19]H. Sies, V.S. Sharov, L.O. Ktotz, K. Briviba, Glutathione peroxidase protects against peroxynitrite-mediated oxidations. A new function for setenoproteins as peroxynitrite reductase, J. Biol. Chem. 272 (1997) 2781227817. [20]Y. Fu, H. Sies, X.G. Lei, Opposite roles of selenium-dependent glutathione peroxidase-1 in superoxide generator diquat- and peroxynitrite-induced apoptosis and signaling, J. Biol. Chem. 276 (2001) 43004-

43009.

H. Budde and L. Flohe / Enzymes in Pathogens as Potential Drug Targets

93

[21]G.E. Arteel, H. Sies, The biochemistry of selenium and the glutathione system, Environmental Toxycology and Pharmacology 10 (2001) 153-158. [22]K.E. Hill, R.F. Burk, Selenoprotein P. In: D.H. Hatfield (Ed.), Selenium. Its Molecular Biology and Role in Human Health. Kluwer Academic Publishers, Boston/ Dordrecht/ London, 2001, pp. 123-135. [23]S.G. Rhee, S.W. Kang, LE. Netto, M.S. Seo, E.R. Stadtman, A family of novel peroxidases, peroxiredoxins, Biofactors 10 (1999) 207-209. [24]A.B. Fisher, C. Dodia, Y. Manevich, J.W. Chen, S.I. Feinstein, Phospholipid hydroperoxides are substrates for non-selenium glutathione peroxidase, J. Biol. Chem. 274 (1999) 21326-21334. [25]B. Hofmann, HJ. Hecht, L. Flohe, Peroxiredoxins, Biol. Chem. 383 (2002) 347-364. [26]A.M. Fairlamb, A. Cerami, Metabolism and functions of trypanothione in the Kinetoplastida, Annu. Rev. Microbiol. 46 (1992) 695-729. [27] K. Koenig, U. Menge, M. Kiess, V. Wray, L. Flohe, Convenient isolation and kinetic mechanism of glutathionylspermidine synthetase from Crithidia fasciculata, J. Biol. Chem. 272 (1997) 11908-11915. [28]E. Tetaud, F. Manai, M.P. Barrett, K. Nadeau, C.T. Walsh, A.H. Fairlamb, Cloning and characterization of the two enzymes responsible for trypanothione biosynthesis in Crithidia fasciculata, J. Biol. Chem. 273 (1998) 19383-19390. [29]M.A. Ouaissi, J.F. Dubremetz, R. Schoneck, R. Fernandez-Gomez, R. Gomez-Corvera, O. Billaut-Mulot, A. Taibi, M. Loyens, A. Tartar, C. Sergheraert, Trypanosoma cruzi: a 52-kDa protein sharing sequence homology with glutathione S-transferase is localized in parasite organelles morphologically resembling reservosomes, Exp. Parasitol. 81 (1995) 453-461. [30]E. Nogoceke, D.U. Gommel, M. Kiess, H.M. Kalisz, L. Flohe, A unique cascade of oxidoreductases catalyses trypanothione-mediated peroxide metabolism in Crithidia fasciculata, Biol. Chem. 378 (1997) 827-836. [31JE. Tetaud, C. Giroud, A.R. Prescott, D.W. Parkin, D. Baltz, N. Biteau, T. Baltz, A.H. Fairlamb, Molecular characterisation of mitochondrial and cytosolic trypanothione-dependent tryparedoxin peroxidases in Trypanosoma brucei, Mol. Biochem. Parasitol. 116 (2001) 171-183. [32]N.M. el-Sayed, C.M. Alarcon, J.C. Beck, V.C. Sheffield, J.E. Donelson, cDNA expressed sequence tags of Trypanosoma brucei rhodesiense provide new insights into the biology of the parasite, Mol. Biochem. Parasitol. 73 (1995) 75-90. [33]S.A. Guerrero, J.A. Lopez, P. Steinert, M. Montemartini, H.M. Kalisz, W. Colli, M. Singh, M.J. Alves, L. Flohe, His-tagged tryparedoxin peroxidase of Trypanosoma cruzi as a tool for drug screening, Appl. Microbiol. Biotechnol. 53 (2000) 410-414. [34]J.A. Lopez, T.U. Carvalho, W. de Souza, L. Flohe, S.A. Guerrero, M. Montemartini, H.M. Kalisz, E. Nogoceke, M. Singh, M.J. Alves, W. Colli, Evidence for a trypanothione-dependent peroxidase system in Trypanosoma cruzi, Free Rad. Biol. Med. 28 (2000) 767-772. [35]S.R. Wilkinson, N.J. Temperton, A. Mondragon, J.M. Kelly, Distinct mitochondrial and cytosolic enzymes mediate trypanothione- dependent peroxide metabolism in Trypanosoma cruzi, J. Biol. Chem. 275 (2000) 8220-8225. [36]M.P. Levick, E. Tetaud, A.H. Fairlamb, J.M. Blackwell, Identification and characterisation of a functional peroxidoxin from Leishmania major, Mol. Biochem. Parasitol. 96 (1998) 125-137. [37]L. Flohe, H. Budde, K. Bruns, H. Castro, J. Clos, B. Hofmann, S. Kansal-Kalavar, D. Krumme, U. Menge, K. Plank-Schumacher, H. Sztajer, J. Wissing, C. Wylegalla, HJ. Hecht, Tryparedoxin peroxidase of Leishmania donovani: molecular cloning, heterologous expression, specificity, and catalytic mechanism, Arch. Biochem. Biophys. 397 (2002) 324-335. [38]S. Krieger, W. Schwarz, M.R. Ariyanayagam, A.H. Fairlamb, R.L. Krauth-Siegel, C. Clayton, Trypanosomes lacking trypanothione reductase are avirulent and show increased sensitivity to oxidative stress, Mol. Microbiol. 35 (2000) 542-552. [39]J. Tovar, M.L. Cunningham, A.C. Smith, S.L. Croft, A.H. Fairlamb, Downregulation of Leishmania donovani trypanothione reductase by heterologous expression of a transdominant mutant homologue: effect on parasite intracellular survival, Proc. Natl. Acad. Sci. U S A 95 (1998) 5311-5316. [40]M. Dormeyer, N. Reckenfelderbaumer, H. Ludemann, R.L. Krauth-Siegel, Trypanothione-dependent synthesis of deoxyribonucleotides by Trypanosoma brucei ribonucleotide reductase, J. Biol. Chem. 276 (2001)10602-10606.

94

H. Budde and L. Flohe / Enzymes in Pathogens as Potential Drug Targets

[41 ]N. ReckenfekJerbaumer, H. LCJdemann, H. Schmidt. D. Steverding, R.L Krauth-Stegel. Identification and functional characterization of thioredoxin from Trypanosoma brucei brucei, J. Bid. Chem. 275 (2000) 75477552. [42]P. Steinert, K. Dittmar, H.M. Kalisz, M. Montemartini, E. Nogoceke, M. Rohde. M. Singh, L Flohe, Cytoplasmic localization of the trypanothione peroxidase system in Crithidia fasciculata, Free Rad. Bid. Med. 26(1999)844-849. [43]S. Wilkinson, D.J. Meyer, M.C. Taylor, E.V. Bromley, M.A. Miles, J.M. Kelly, The Trypanosoma cruzi enzyme TcGPxl is a glycosomal peroxidase and can be linked to trypanothione reduction by gkrtathione or tryparedoxin, J. Bid. Chem. (2002) in press. [44]B. Gamain, G. Langsley, M.N. Fourmaux, J.P. Touzel, D. Camus, D. Dive, C. Skxnianny, Molecular characterization of the glutathione peroxidase gene of the human malaria parasite Plasmodium fakaparum, Mol. Biochem. Parasitol. 78 (1996) 237-248. [45]H. Sztajer, B. Gamain, K.D. Aumann, C. Stomianny, K. Becker, R. Brigelius-Ftohe, L Flohe, The putative glutathione peroxidase gene of Plasmodium faltiparum codes for a thioredoxin peroxidase, J. Biol. Chem. 276(2001)7397-7403. [46]S. Rahlfs, K. Becker, Thioredoxin peroxidases of the malarial parasite Plasmodium fataparum, Eur. J. Biochem. 268 (2001) 1404-1409. [47]S. Rahlfs, M. Fischer, K. Becker, Plasmodium falciparum possesses a classical glutaredoxin and a second, glutaredoxin-like protein with a PICOT homology domain, J. Biol. Chem. 276 (2001) 37133-37140. [48]M.R. Ariyanayagam, A.M. Fairiamb, Entamoeba histolytica lacks trypanothione metabolism, Mol. Biochem. Parasitol. 103 (1999) 61-69. [49]R.N. Ondarza, E.M. Tamayo, G. Hurtado, E. Hernandez, A. Iturbe, Isolation and purification of glutathionyrtspermidine and trypanothione from Entamoeba histolytica. Arch. Med. Res. 28 (1997) 73-75. [50]R.N. Ondarza, A. Iturbe, G. Hurtado, E. Tamayo, M. Ondarza. E. Hernandez, Entamoeba histolytica: a eukarvote with trypanothione metabolism instead of glutathione metabolism, Biotechnol. Appl. Biochem. 30 (1999)47-52. [51]l. Bruchhaus, S. Richter, E. Tannich, Removal of hydrogen peroxide by the 29 kOa protein of Entamoeba histolytica, Biochem. J. 326 (1997) 785-789. [52]LB. Pode, H.Z. Chae, B.M. Flores, S.L Reed, S.G. Rhee, B.E. Torian, Peroxidase activity of a TSA-like anttoxidant protein from a pathogenic amoeba, Free Rad. Biol. Med. 23 (1997) 955-959. [53]F.S. Jacobson, R.W. Morgan, M.F. Christman, B.N. Ames, An alkyl hydroperoxide reductase from Salmonella typhimurium involved in the defense of ONA against oxidative damage. Purification and properties, J. Biol. Chem. 264 (1989) 1488-1496. [54]G. Storz, F.S. Jacobson, LA. Tartaglia, R.W. Morgan, L.A. Silveira. B.N. Ames, An alkyl hydroperoxide reductase induced by oxidative stress in Salmonella typhimurium and Escherichia coli: genetic characterization and cloning of ahp, J. Bacteriol. 171 (1989) 2049-2055. [55]S. Dhandayuthapani, Y. Zhang, M.H. Mudd, V. Deretic, Oxidative stress response and its rote in sensitivity to isoniazid in mycobacteria: characterization and inducibility of ahpC by peroxides in Mycobacterium smegmatis and lack of expression in M. aurum and M. tuberculosis, J. Bacteriol. 178 (1996) 3641-3649. [56]D.R. Sherman, K. Mdluli, M.J. Mickey, T.M. Arain, S.L Morris, C.E. Barry, 3rd, C.K. Stover, Compensatory ahpC gene expression in isoniazid-resistant Mycobacterium tuberculosis, Science 272 (1996) 1641-1643. [57JL.M. Baker, A. Raudonikiene, P.S. Hoffman, LB. Poote, Essential thioredoxin-dependent peroxiredoxin system from Helicobacter pylori: genetic and kinetic characterization, J. Bacteriol. 183 (2001) 1961-1973. [58]D.R. Sherman, K. Mdluli, M.J. Mickey, C.E. Barry, 3rd, C.K. Stover, AhpC, oxidative stress and drug resistance in Mycobacterium tuberculosis, Biof actors 10 (1999) 211 -217. [59]R. Bryk, C.D. Lima, H. Erdjument-Bromage, P. Tempst, C. Nathan, Metabolic enzymes of mycobacteria linked to antioxidant defense by a thioredoxin-like protein, Science 295 (2002) 1073-1077. [60]G.L Newton, K. Arnold, M.S. Price, C. Sherrill, S.B. Delcardayre, Y. Aharonowitz, G. Cohen, J. Davtes. R.C. Fahey, C. Davis, Distribution of thiols in microorganisms: mycothid is a major thid in most actinomycetes, J. Bacteriol. 178 (1996) 1990-1995. [61]E.M. Jacoby, I. Schlichting, C.B. Lantwin, W. Kabsch, R.L. Krauth-Stegel, Crystal structure of the Trypanosoma cruzi trypanothione reductase mepacrine complex, Proteins 24 (1996) 73-80.

H. Budde and L. Flohe / Enzymes in Pathogens as Potential Drug Targets

95

[62]C.S. Bond, Y. Zhang, M. Berriman, M.L Cunningham, A.M. Fairlamb, W.N. Hunter, Crystal structure of Trypanosoma cruzi trypanothione reductase in complex with trypanothione, and the structure-based discovery of new natural product inhibitors, Structure Fold Des. 7 (1999) 81-89. [63JM.S. Alphey, C.S. Bond, E. Tetaud, A.H. Fairlamb, W.N. Hunter, The structure of reduced tryparedoxin peroxidase reveals a decamer and insight into reactivity of 2Cys-peroxiredoxins, J. Mol. Bio). 300 (2000) 903-916. [64JJ.R. Harris, E. Schroder, M.N. Isupov, D. Scheffler, P. Kristensen, J.A. Ltttlechild, A.A. Vagin, U. Meissner, Comparison of the decameric structure of peroxiredoxin-ll by transmission electron microscopy and X-ray crystallography, Biochim. Biophys. Ada 1547 (2001) 221-234. [65]B. Hofmann, H. Budde, K. Bruns, S.A. Guerrero, H.M. Kalisz, U. Menge, M. Montemartini, E. Nogoceke, P. Steinert, J.B. Wissing, L. Flohe, H.J. Hecht, Structures of tryparedoxins revealing interaction with trypanothione, Biol. Chem. 382 (2001) 459-471.

96

Thiol Metabolism and Redox Regulation of Cellular Functions A. Pompella et al. (Eds.) IOS Press, 2002

Is there a Role of Glutathione Peroxidases in Signaling and Differentiation? Regina BRIGELIUS-FLOH£l) and Leopold FLOHE2) German Institute of Human Nutrition, Arthur-Scheunert-Allee 114-116, D-14558 Bergholz-Rehbrucke, Germany; 2) Dept. of Biochemistry, Technical University of Braunschweig, Mascheroder Weg 1, D-38124 Braunschweig, Germany.

l)

1. Introduction Mammalian organisms contain five distinct isozymes of glutathione peroxidase (GPx), four of them being selenoproteins (for review see [1]). The enzyme family is characterized by a typical catalytic triad, in which the sulfur or selenium of a (seleno)cysteine residue is activated by hydrogen bonding to a tryptophan and a glutamine residue [2]. In the selenium-containing examples (cGPx or GPx-1, GI-GPx or GPx-2, pGPx or GPx-3, PHGPx or GPx-4), the selencysteine residue is oxidized by hydroperoxides with rate constants k'i up to 108 M~V. The corresponding k'j of the sulfur homolog GPx-5 is unknown, but investigations on sulfur-containing muteins of cGPx [3] and PHGPx [2] and on a protozoal homolog [4] suggest that sulfur-mediated catalysis in general falls short by at least two orders of magnitude when compared to selenium catalysis. The designation "glutathione peroxidase" of the family is based on the pronounced specificity of cGPx for glutathione as reducing substrate [5,6], but must not uncritically be considered to describe the biological role of all family members. The extracellular GPx-3 equally accepts thioredoxin and glutaredoxin [7], PHGPx can obviously be reduced by a variety of protein thiols [8-10] and the homolog of Plasmodium falciparum is specific for thioredoxin [4]. Inversely, enzymes of other families reportedly display glutathione peroxidase activity, e.g., glutathione-Stransferase [11] and the type VI peroxiredoxin [12]. The glutathione peroxidases have for long been appreciated as antioxidant devices which, in concerted actions with superoxide dismutases, just prevent oxidative tissue damage. In this scenario, the ubiquituous cGPx was supposed to protect the cytosol and the mitochondrial matrix [13]. PHGPx, the phospholipid hydroperoxide glutathione peroxidase, was regarded as pivotal for biomembrane protection [14], the gastrointestinal GI-GPx was discussed to shield the organism against food-derived peroxides [15] and GPx-3 was supposed to protect the extracellular space [16]. However, numerous observations reported over the past decade do not comply with the simplistic view that the multiplicity of glutathione peroxidases means little else than compartmented antioxidant defence. The aim of this article therefore is to compile the hints pointing to distinct functions of the individual types of glutathione peroxidases. 2. Antioxidant Defence versus Redox Regulation Also in biology, definitions only make sense if they discriminate between phenomena. The term oxidative stress has been created to describe pathological conditions in which an organism or tissue is damaged by an unbalanced production of oxidants such as HaCh, other

R. Brigelius-Flohe and L. Flohe / Glutathione Peroxidases in Signaling

97

hydroperoxides and oxygen- or nitrogen-centered radicals. Such situation results from massive activation of NADPH oxidase in phagocytes, as in infectious or inflammatory diseases, from exposure to hyperbaric oxygen, that leads to mitochondrial (V" production, from ischemia/reperfusion or from poisoning with redox-cycling xenobiotics. If the term oxidative stress is diluted in a way to comprise all kind of oxidative events going on in an undisturbed metabolism of an aerobic organism, it becomes meaningless. The same consideration pertains to the fashionable term "antioxidant defence". Sensu stricto it should define a mechanism to prevent pathologies resulting from oxidative stress and not, e.g., any reduction of a hydroperoxide at physiological level. As is to be demonstrated, this is not a semantic distinction. The NADPH oxidase, that releases superoxide in phagocytes engaged in host defense (Fig. 1), is similarly activated in other cells upon stimulation with cytokines, growth factors or hormones (for review see [17]). The alternate stimuli comprise cytokines like TNF or IL-1 that are typically involved in host defense too, growth factors like GMCSF that indirectly contribute to host defense by recruiting immune competent cells, but also hormones such as insulin that are evidently unrelated to the need to fight pathogens by a cocktail of toxic oxidants. As a rule the 'oxidative burst' associated with such physiological stimuli is by far smaller than the response of phagocytes to pathogens, and the H2(>2 thus formed must be considered as a second messenger that is indispensible for normal cell proliferation and function. As is evident for Ha02, lipoxygenase products can only be rated as toxic when formed in excess; their pharmacodynamic profiles are too diverse to be discussed here. It may suffice to state that mammalian physiology can hardly be envisaged without leukotrienes and prostaglandins. In consequence, the role of enzymes acting on HaOa and organic hydroperoxides cannot generally be seen in utmost elimination of such oxidants; depending on tissue distribution, compartmentation and specificities they rather will cover the whole range of peroxide detoxification up to regulation of physiological processes. A typical example for antioxidant defense enzymes is cGPx. It is the most abundant and ubiquituously distributed one of the glutathione peroxidases. Nevertheless, cGPx (-/-) mice develop normally and grow even faster than wild-type mice. They don't show any obvious phenotype [21,22]. Apparently, the enzyme is not needed in an unstressed organism. However, when the cGPx (-/-) mice were challenged with redox-cycling herbicides [22,23], with bacterial lipopolysaccharides to mimic septicemia [24] or with viral infections [25] they died faster. In this respect, the k.o. mice resemble human subject deficient in G6PDH, an enzyme that keeps GPx in function by supplying the NADPH for GSH regeneration. Also such subjects do not show any phenotype if they are not exposed to pro-oxidative redox-cycling antibiotics or xenobiotics contained in fava beans [26]. These characteristics of an antioxidant device are by no means shared by all glutathione peroxidases. A sharply contrasting example is PHGPx. Its tissue distribution is quite unusual in being highest in endocrine organs (reviewed in [27]), it proved to be essential for spermatogenesis (see below), and targeted disruption of the gene in mice led to early embryonic lethality (M. Brielmeier, personal communication). The latter observation reveals that PHGPx is of vital importance, which is not yet understood. Nevertheless, the idea that the enzyme might be essential for normal development, because it prevents the embryos from getting rancid, sounds naive. In fact, amongst the mammalian peroxidases, cGPx appears to be quite unique in fighting a generalized oxidative stress. The model experiments with the cGPx (-/-) mice quoted above were performed with selenium-deficient and selenium-adequate mice. Surprisingly, the selenium status did not significantly affect the outcome, which implies that neither the remaining three selenoperoxidases nor any of the other selenoproteins can

98

R. Brigelius-Flohe and L Flohe / Glutathione Peroxidases in Signaling

efficiently substitute for cGPx in antioxidant defense. Inversely spoken, the roles of GPx-2, 3 and 4 should be searched for in different biological contexts.

Rg. 1. Routes to activation of NADPH oxkJase. The silent enzyme consisting of gp91phox and p22phox needs to be activated for host defense by recruitment of p47phox and p67phox [18]. Prerequisites are polyphosphorylation of p47phox by PKC and activation of the p67 carrier Rac by phosphatidylinosttol-3-phosphate (PIP3) [19,20]. Activation of phosphatidylinositol-3-kinase (PI3K) upon stimulation, e.g., of the PDGF receptor, is implicated in associated H2O2 formation. Analogous responses can be expected by any kind of receptor stimulation leeding to activation of PI3K and PKC.

3. Regulatory Phenomena attributed to Glutathione Peroxidases 3.1. Eicosanoid synthesis cGPx plus GSH was shown to completely block prostaglandin biosynthesis in sheep seminal vesicles already in 1972 [28]. Analogous observations were made with 5-

R. Brigelius-Flohe and L. Flohe / Glutathione Peroxidases in Signaling

99

lipoxygenase, the key enzyme of leukotriene biosynthesis [29]. In both cases, the inhibition could be overcome by addition of organic hydroperoxides. PHGPx was shown to have the same effect on 15-lipoxygenase [30]. Taking advantage of the differential response of cGPx and PHGPx to selenium deprivation/repletion (reviewed in [27]), Weitzel and Wendel [31] demonstrated that in vivo PHGPx is more relevant to leukotriene biosynthesis. h Overexpression of PHGPx in rat basophilic leukemia cells corroborates a role of PHGPx in the regulation of leukotriene [32] and prostaglandin [33] synthesis. A significant part of overexpressed PHGPx was found in, or attached to the nucleus. This appears logical since prostaglandin H synthase 1 and 2 localize in the nuclear envelope and endoplasmic reticulum to which phospholipase A2 and 5-lipoxygenase have to be translocated from the cytosol [34,35]. If also GI-GPx would regulate prostaglandin and leukotriene synthesis in intestinal cells, its distinct localization in nucleus-associated structures appears in a new light (see below). Apart from affecting the activity of lipoxygenases, glutathione peroxidases may also affect leukotriene biosynthesis by reducing, e. g., the primary product of 5-lipoxygenase, 15-HPETE, to 5 HETE thereby draining away the supply for synthesis of the pharmacologically active leukotriene 64 and peptidoleukotrienes [36,37]. 3.2. NFKB activation Overexpression of cGPx leads to a dampened NFKB activation in response to TNF-a [38,39]. The most upstream point of interference with the TNF-a signalling cascade so far identified is 1KB, the inhibitory component of the inactive cytosolic NFKB complex. cGPx inhibits 1KB phosphorylation, thereby preventing ubiquitination and degradation of 1KB and the consecutive nuclear translocation of activated NFKB and gene activation. Analogous observations were made with ECV and rabbit aortic smooth muscle cells overexpressing PHGPx [40,41]. In ECV cells interleukinl-induced NFKB activation (Fig. 2) was completely abrogated [40]. Interestingly, large variations of cGPx activity achieved by selenium-deprivation/repletion only marginally affected IL-1-induced NFKB activation, while a moderate change in PHGPx activity had dramatic effects. Whether PHGPx is more relevant to NFKB activation in general, remains to be demonstrated. 3.3. Apoptosis Programmed cell death has amply been documented to be facilitated by pro-oxidative conditions or to be triggered by subtoxic levels of hydroperoxides [49,50], although also redox-independent mechanisms are known. Accordingly, Overexpression of cGPx in T lymphocytes inhibited apoptosis induced by interleukin-3 withdrawal [51] or HIV infection [52]. Similarly, PHGPx appears to inhibit apoptosis when triggered by e. g. deoxyglucose, staurosporine, UV irradiation, or hydroperoxide [41,53,54], whereas the Fas-mediated pathway was not affected [53]. It also appears revealing that overexpressed PHGPx was particularly effective when targeted to the mitochondria [53], since the mitochondiial death pathway is known to involve generation of oxidants. The physiological balance between proliferation and apoptosis is particularly relevant in steadily regenerating tissue such as the intestinal epithelium. In human colon GI-GPx displays a gradient declining from the proliferating crypts to the top of the villi prone to apoptosis [55]. Whereas GI-GPx in cells from the crypt ground is spread throughout the cytoplasm it is organized in distinct structures at the apical pole of the nuclei in cells of the luminal site of the crypt. The structures remind of an association with the Golgi system. The strikingly different distribution in proliferating, less differentiated crypt ground cells and in differentiated luminal cells prone to be eliminated by apoptosis might certainly reflect distinct physiological functions) at different developmental stages of colon

100

R. Brigelius-Flohe and L. Flohe / Glutathione Peroxidases in Signaling

cells. GI-GPx (-/-) mice were unsymptomatic. Only when cross-bread with GPx-1 (-/-) mice they developed colitis [56]. Whether these findings can be regarded to indicate a role of GI-GPx in sustaining the delicate balance of proliferation and apoptosis remains elusive. Certainly other roles of GI-GPx have to be considered [27,55,57] (see also Budde & Flohe, this volume).

Fig. 2. Co-production of oxldants by IL-1. IL-1 binds to the IL-1 receptor type I (IL-1RI) which heterodimerizes with the IL-1 receptor accessory protein (IL-1RAcP) [42]. Then the IL-1Rassociated kinase (IRAK) is recruited [43] and associated via the adapter protein MyD88 [44]. Signaling pathways include the production of *O2~ and HjO2 [45] by the cell type-specific activation of NADPH oxidase or 5-lipoxygenase [46]. The mechanisms involved are not dear. H2O2 and lipoxygenase products facilitate signaling by the phosphoryiation cascade upstream of kB, but may inhibit binding of p50 to DNA. Overexpression of PHGPx dampens IL-1-induced NFicB activation (see text). However, oxidation [47] and glutathiolatJon [48] of pSO inhibits binding of the transcription factor to DNA.

4. PHGPx as Protein Thiol Peroxidase and Structural Protein Enzymatically inactive PHGPx protein was identified as the main component of the keratin-like material that embeds the helix of mitochondria in the midpiece of spermatozoa

R. Brigelius-Flohe and L. Flohe /Glutathione Peroxidases in Signaling

101

[9]. As such it appears indispensible for structural and functional integrity of spermatozoa. Active PHGPx can be recovered from this material only by drastic reductive procedures such as treatment with 0.1M DTT for hours indicating that the peroxidase had been oxidatively cross-linked with itself and/or other proteins. The use, by PHGPx, of protein thiols in solubilized capsule proteins as alternate substrates was also demonstrated in vitro [58]. Likewise, PHGPx was reported to react with nuclear proteins [8] and a splicevariant of PHGPx has recently been detected in nuclei of spermatogenic cells, where it is supposed to be covalently bound to chromatin thereby regulating cell division [59].

5. Potential Regulatory Mechanisms The frequently heard statement that "ROS" (reactive oxygen species) regulate many signaling cascades or differentiation processes is too imprecise to be satisfactory. In fact, the most reactive ones, -OH or RO, would be too promiscuous to meet any requirement for specificity. They also tend to irreversibly modify proteins and thus are rather relevant to oxidative tissue damage. ROS suited for signalling events have to be produced in a regulated and compartmented manner and their scope of reactivities should allow reversible modifications of protein targets. The molecular events that are involved in specific redox regulation by such components are numerous: i) Enzymes may directly be activated by H2O2 or organic hydroperoxides according to eq. (1).

Inactive enzyme + ROOM •> active enzyme + ROM

(1).

This mechanism likely explains the quoted silencing of cyclooxygenases and lipoxygenases by glutathione peroxidases plus GSH. These usually dormant enzymes require a certain peroxide tone to become activated. Elimination of ROOH by any kind of GPx will therefore put them to rest again. Some specificity of the activation process appears to be achieved by the lipophilicity of the enzymes' active sites. Accordingly, hydroperoxy fatty acids are more effective than HaC^ in lipoxygenase activation. The in vivo superiority of PHGPx in silencing 5-lipoxygenase may be due to the preference of PHGPx for lipophilic hydroperoxides or to microcompartmentation. An oxidation of critical cysteine residues to sulfenic acids has been implicated in "redox priming" of the insulin receptor receptor kinase, which means that autophosphorylation of the activation loop is facilitated and the substrate specificity is altered. Upon oxidative modification autophosphorylation is achieved by phosphocreatine rather than by ATP [60]. Inversely, ROOH may directly inactivate enzymes. This effect is incidentally also observed with lipoxygenases, which are known to be product-inactivated. Whether the implicated overoxidation of the reaction centers is achieved by hydroperoxides or free radicals locally derived therefrom, is not clear. An inactivation by H2O2 or other peroxides, in particular by peroxovanadated, is also observed with many protein phosphatases explaining the often observed increase in protein phosphorylation. All protein tyrosine phosphatases (PTPs) contain an essential cysteine residue in the active site motif, His-Cys-X-X-Gly-X-X-Arg-Ser/Thr (reviewed in [17]). The neighbouring basic amino acids lower the pKa of the cysteine residue to 4.7-5.4 (usually around 8.5) [61]. The resulting thiolate anion at physiological pH is the target for the attack of H2O2. The resulting sulfenic acid is re-reduced most efficiently by thioredoxin. PP2B, a protein Ser/Thr phosphatase, contains the active site motif Cys-XX-Cys and is inactivated by ^62 and the vicinal thiol modifying agent phenyarsine oxide. Oxidative

102

R. Brigelius-Flohe and L Flohe I Glutathione Peroxidases in Signaling

inactivation is reportedly due to the formation of a disulfide bond between Cys Cys256 [62].

and

ii) A shift in the GSH/GSSG balance may favor the formation of mixed disulfides between GSH and proteins according to (2).

Prot -S" + GSSG <-> Prot-S-SG + GS"

(2)

Specificity could be achieved by the microenvironment of the cysteine residue in the protein that has to enforce thiol dissociation and to allow sterical access of GSSG. Such glutathiolation has been documented for an increasing number of proteins reviewed in [63]. In most of these examples glutathiolation could equally be achieved when GSSG was replaced by nitrosoglutathione [63]. Reversal of protein glutathiolation may occur directly as indicated by e.g. (2) but appears to be facilitated by glutaredoxin [63]. Alternatively, it might be proceeded by modification of the protein bound GSH. As an early example of the latter case the regulation of fructose 1,6 diphosphatase could be quoted. Here a cysteamine was found bound to the protein [64]. Previous modification of the glutathionyl residue offers the chance of independent regulation of the reverse reaction. iii) Protein thiolation could equally be achieved differently according to eq. (3) and (4). Prot-S" + ROOM -» Prot - SOH + ROM

(3)

Prot-SOH + R'SH -» Prot-S-SR' + H2O

(4)

This sequence of reactions has been documented to represent part of the catalytic principle of peroxiredoxins [65-67] and, with or without replacement of S by Se, of GPx-type enzymes [1]. In these cases, the regeneration of the ground state enzymes (prot-S (Se)~) from Prot-S(Se)-SR' is achieved by thiol disulfide exchange with reducing substrates, such as GSH (equation 5) or reduced thioredoxin, respectively.

Prot-S-SR1 + GSH <-* Prot-S" + GSSR + H+

(5)

There is no reason to assume that the reaction (3) to (5) are restricted to the two peroxidase families. In fact, every exposed and dissociated cysteine or selenocysteine residue can be oxidized by ROOH to a sulfenic or selenenic acid derivative that will readily react with thiols, be it GSH or an SH group of a suitable protein. These reactions offer ample opportunity for specific regulation if the mixed disulfides thereby formed are reasonably stable and alter the activity of the protein(s) involved. Oxidation of cysteines to sulfenic acids, with of without consecutive formation of intramolecular disulfide bonds, is being discussed as the molecular basis of oxidative inhibition of transcription factors such as AP-1, NF KB (see fig. 2), nuclear factor-1, Sp-1, hypoxiainducible factor la, and p53 [63]. iv) A special case of the above principle (iv) would be the use of GPx-type enzymes as a thiol modifying agent. The selenocysteine in their active site is particularly prone to become oxidized to a selenenic acid (see chapter 1). In case of PHGPx, the reaction of the oxidized enzyme with itself and other proteins according to (6) and (7) n HS-PHGPx-SeOH -* (S-GPx-Se)n + n H2O (6) PHGPx-SeOH + Prot-SH -» PHGPx-Se-S-Prot + H2O (7) is evident from its transformation during spermatogenesis. It should be stressed that, in this particular situation, PHGPx acts precisely as the opposite of an antioxidant device. It makes use of peroxides to build up a structural element that is indispensible for appropriate function of the cell. It would be surprising if nature had preserved the

R. Brigelius-Flohe and L Flohe / Glutathione Peroxidases in Signaling

103

underlying chemical principle just for the final step of mammalian spermatogenesis. In view of the increasing number or poorly explained phenomena suggesting a specific involvement of PHGPx in regulatory events, equation (7) merits attention as possible regulatory principle. It may equally be considered for GPx-3 that can still be regarded as an orphan enzyme in being devoid of adequate substrate supply and proven function. v) Finally, hydroperoxides via GPx-3 or peroxiredoxins will affect the redox state of the thioredoxin system that has been amply documented to regulate numerous cellular events. A particular interesting case is the regulation of the apoptosis signal-regulating kinase-1 (ASK1). ASK1 is associated with thioredoxin (Trx), which prevents the interaction with TRAF2 upon TNF-stimulation. Oxidation of Trx(SH)2 to TrxS2 leads to a dissociation of Trx, making the way free for downstream signals [68,69].

6. Conclusions Ever since the discovery of cGPx in the late fifties [5], the vision has been blurred by focussing almost exclusively on antioxidant defense by glutathione peroxidases. Emerging evidence points to distinct roles of the remaining selenoperoxidases, particularly in signaling and cellular differentiation. The molecular mechanisms are mostly unclear. But the chemical principles, which became knowm from the analyses of catalytic mechanisms and model reactions, provide numerous hypotheses that can now be scrutinized. It is anticipated that redox regulation, directly or indirectly mediated by glutathione peroxidases and/or peroxiredoxins, will soon complement research on protein kinases afrd phosphatases that up to now dominates cell biology.

Acknowledgements This work was supported by the Deutsche 'Selenproteins', grants to R.B.F and L.F.).

Forschungsgemeinschaft

(Priority

Program

References [1] L. Flohe and R. Brigelius-Flohe, Selenoproteins of the glutathione system. In: D.L. Hatfield (ed.), Selenium. Its molecular biology and role in human health. Kluwer Academic Publishers, Boston, 2001, pp. 157-178. [2] M. Maiorino, K.D. Aumann, R. Brigelius-Flohe, D. Doria, J. van den Heuvel, J. McCarthy, A. Roveri, F. Ursini, L. Flohe, Probing the presumed catalytic triad of selenium-containing peroxidases by mutational analysis of phospholipid hydroperoxide glutathione peroxidase (PHGPx), Biol. Chem. Hoppe Seyler. 376 (1995)651-660. [3] C. Rocher, J.L. Lalanne, J. Chaudiere, Purification and properties of a recombinant sulfur analog of murine selenium-glutathione peroxidase, Eur. J. Biochem. 205 (1992) 955-960. [4] H. Sztajer, B. Gamain, K.-D. Aumann, C. Slomianny, K. Becker, R. Brigelius-Flohe, L. Flohe, The putative glutathione peroxidase gene of Plasmodium falciparum codes for a thioredoxin peroxidase, J. Biol. Chem. 276(2001)7397-7403. [5] G.C. Mills, Hemoglobin Catabolism I. Glutathione peroxidase, an erythrocyte enzyme which protects hemoglobin from oxidative breakdown, J. Biol. Chem. 229 (1957) 189-197. [6] L. Ftohe, W. Giinzler, G. Jung, E. Schaich, F. Schneider, [Glutathione peroxidase. II. Substrate specificity and inhibitory effects of substrate analogues], Hoppe Seylers Z. Physiol. Chem. 352 (1971) 159-169. [7] M. Bjornstedt, J. Xue, W. Huang, B. Akesson, A. Holmgren, The thioredoxin and glutaredoxin systems are efficient electron donors to human plasma glutathione peroxidase, J. Biol. Chem. 269 (1994) 29382-29384.

104

R. Brigelius-Flohe and L. Flohe / Glutathione Peroxidases in Signaling

[6] C. Godeas, F. Tramer, F. Micali, A. Roveri, M. Maiorino, C. Nisii, G. Sandri. E. PanfiN, Phosphofipid hydroperoxide glutathione peroxkjase (PHGPx) in rat testis nudei is bound to chromatin, Btochem. Mol. Med. 59(1996)118-124. [9] F. Ursini, S. Heim, M. Ktess, M. Maiorino, A. Roveri, J. Wissing, L Flohe, Dual function of the setenoprotein PHGPx during sperm maturation [see comments], Science. 285 (1999) 1393-1396. [10]M. Maiorino, L Flohe, A. Roveri, P. Steinert, J.B. Wissing, F. Ursini, Selenium and reproduction, BioFactors. 10 (1999) 251-256. [11]B. Ketterer and D.J. Meyer, Glutathione transferases: a possible role in the detoxication and repair of DMA and lipid hydroperoxides, Mutat. Res. 214 (1989) 33-40. [12JA.B. Fisher, C. Dodia, Y. Manevich, J.W. Chen, S.I. Feinstein. Phospholipid hydroperoxides are substrates for non-selenium glutathione peroxidase, J. Biol. Chem. 274 (1999) 21326-21334. [13]L Flohe and W. Schlegel, [Glutathione peroxidase. IV. Intracellular distribution of the glutathione peroxidase system in the rat liver], Hoppe Seylers Z. Physiol. Chem. 352 (1971) 1401-1410. [14]F. Ursini, M. Maiorino, M. Valente, L. Fern, C. Gregolin, Purification from pig liver of a protein which protects liposomes and biomembranes from peroxidative degradation and exhibits glutathione peroxidase activity on phosphatidylcholine hydroperoxides, Biochim. Biophys. Ada 710 (1982) 197-211. [15]F.F. Chu, J.H. Ooroshow, R.S. Esworthy, Expression, characterization, and tissue distribution of a new cellular selenium-dependent glutathione peroxidase, GSHPx-GI, J. Biol. Chem. 268 (1993) 2571-2576. [16]K. Takahashi, N. Avissar, J. Whitin, H. Cohen, Purification and characterization of human plasma glutathione peroxidase: a selenoglycoprotein distinct from the known cellular enzyme, Arch. Biochem. Biophys. 256 (1987) 677-686. [17]S.G. Rhee, Y.S. Bae, S.R. Lee, J. Kwon, Hydrogen peroxide: a key messenger that modulates protein phosphorylation through cysteine oxidation, Sci STKE. 2000 (2000) E1. [18JB.M. Babior, NADPH oxktose: an update, Blood. 93 (1999) 1464-1476. [19]Y. Gorzalczany, N. Sigal, M. Itan, O. Lotan, E. Pick, Targeting of Rac1 to the phagocyte membrane is sufficient for the induction of NADPH oxidase assembly, J. Biol. Chem. 275 (2000) 40073-40081. [20JY.S. Bae, J.Y. Sung, O.S. Kim, Y.J. Kim, K.C. Hur, A. Kazlauskas, S.G. Rhee, Platelet-derived growth factor-induced H(2)O(2) production requires the activation of phosphatidytinositol 3-kinase, J. Biol. Chem. 275(2000)10527-10531. [21]Y.S. Ho, J.L Magnenat, R.T. Bronson, J. Cao, M. Gargano, M. Sugawara, C.D. Funk, Mice deficient in cellular glutathione peroxidase develop normally and show no increased sensitivity to hyperoxia, J. Biol. Chem. 272 (1997) 16644-16651. [22JX.G. Lei, Glutathione peroxidase-1 gene knockout on body antioxidant defense in mice, BioFactors. 14 (2001)93-99. [23]W.H. Cheng, Y.S. Ho, B.A. Valentine, D.A. Ross, G.F. Combs. Jr., X.G. Lei, Cellular glutathione peroxidase is the mediator of body selenium to protect against paraquat lethality in transgenic mice, J. Nutr. 128 (1998) 1070-1076. [24]H. Jaeschke, Y.S. Ho, M.A. Fisher, J.A. Lawson, A. Farhood, Glutathione peroxidase-deficient mice are more susceptible to neutrophil-mediated hepatic parenchyma! cell injury during endotoxemia: importance of an intracellular oxkJant stress, Hepatotogy. 29 (1999) 443-450. [25]M.A. Beck, Selenium as an antiviral agent. In: D. Hatfiekj (ed.), Selenium. Its molecular biology and role in human health. Kluwer Academic Publishers, Boston, Dordrecht. London, 2001, pp. 235-245. [26]E. Beutler, Red cell enzyme deficiencies as non-disease, Biomed. Biochim. Acta. 42 (1983) S234-S241. [27]R. Brigelius-Flohe, Tissue-specific functions of individual glutathione peroxidases, Free Radic. Biol. Med. 27(1999)951-965. [28]W.L. Smith and W.E. Lands, Oxygenation of polyunsaturated fatty acids during prostaglandin biosynthesis by sheep vesicular gland, Biochemistry. 11 (1972) 3276-3285. [29]M. Haurand and L Flohe, Kinetic studies on arachidonate 5-lipoxygenase from rat basophilic leukemia cells, Biol. Chem. Hoppe Seyler. 369 (1988) 133-142. [30]K. Schnurr, J. Belkner, F. Ursini, T. Schewe, H. Kuhn, The setenoenzyme phospholipid hydroperoxide glutathione peroxidase controls the activity of the 15-lipoxygenase with complex substrates and preserves the specificity of the oxygenation products, J. Biol. Chem. 271 (1996) 4653-4658.

R. Brigelius-Flohe and L Flohe / Glutathione Pemxidases in Signaling

105

[31 ]F. Weitzel and A. Wendel, Selenoenzymes regulate the activity of leukocyte 5-lipoxygenase via the peroxide tone, J. Biol. Chem. 268 (1993) 6288-6292. [32]H. Imai, K. Narashima, M. Arai, H. Sakamoto, N. Chiba, Y. Nakagawa, Suppression of leukotriene formation in RBL-2H3 cells that overexpressed phospholipid hydroperoxide glutathione peroxidase, J. Biol. Chem. 273 (1998) 1990-1997. [33]H. Sakamoto, H. Imai, Y. Nakagawa, Involvement of phospholipid hydroperoxide glutathione peroxidase in the modulation of prostaglandin 02 synthesis, J. Biol. Chem. 275 (2000) 40028-40035. [34]R. Malaviya and B.A. Jakschik, Reversible translocation of 5-lipoxygenase in mast cells upon IgE/antigen stimulation, J. Biol. Chem. 268 (1993) 4939-4944. [35]S. Glover, M.S. de Carvalho, T. Bayburt, M. Jonas, E. Chi, C.C. Leslie, M.H. Gelb, Translocation of the 85kDa phospholipase A2 from cytosol to the nuclear envelope in rat basophilic leukemia cells stimulated with calcium ionophore or IgE/antigen, J. Biol. Chem. 270 (1995) 15359-15367. [36]R.W. Bryant, T.C. Simon, J.M. Bailey, Hydroperoxy fatty acid formation in selenium deficient rat platelets: coupling of glutathione peroxidase to the lipoxygenase pathway, Biochem. Biophys. Res. Commun. 117 (1983)183-189. [37]W.A. Gunzler, M. Lemke, G. Loschen, W. Martin, H. Sohngen, L. Flohe, Investigation on the hypothetical role of glutathione peroxidase in leukotriene formation by rat white blood cells. In: R. Rotilio (ed.), Superoxide and superoxide dismutase in chemistry, biology and medicine. Elsevier Science Publishers B.V., 1986, pp. 382-384. [38]C. Kretz-Remy, P. Mehlen, M.E. Mirault, A.P. Arrigo, Inhibition of I kappa B-alpha phosphorylation and degradation and subsequent NF-kappa B activation by glutathione peroxidase overexpression, J. Cell. Biol. 133(1996)1083-1093. [39]C. Kretz-Remy and A.P. Arrigo, Selenium: a key element that controls NF-kappa B activation and I kappa B alpha half life, BioFactors. 14 (2001) 117-125. [40]R. Brigelius-Flohe, B. Friedrichs, S. Maurer, M. Schultz, R. Streicher, lnterleukin-1-induced nuclear factor kappa B activation is inhibited by overexpression of phospholipid hydroperoxide glutathione peroxidase in a human endothelial cell line, Biochem J. 328 (1997) 199-203. [41]R. Brigelius-Ftohe, S. Maurer, K. Lotzer, G. B6I, H. Kallionpaa, P. Lehtolainen, H. Viita, S. Yla-Herttuala, Overexpression of PHGPx inhibits hydroperoxide-induced oxidation, NFkappaB activation and apoptosis and affects oxLDL-mediated proliferation of rabbit aortic smooth muscle cells, Atherosclerosis. 152 (2000) 307-316. [42JS.A. Greenfeder, P. Nunes, L. Kwee, M. Labow, R.A. Chizzonite, G. Ju, Molecular cloning and characterization of a second subunit of the interleukin 1 receptor complex, J. Biol. Chem. 270 (1995) 13757-13765. [43]M. Martin, G.F. B6I, A. Eriksson, K. Resch, R. Brigelius-Flohe, lnterleukin-1-induced activation of a protein kinase co-precipitating with the type I interleukin-1 receptor in T cells, Eur. J. Immunol. 24 (1994) 15661571. [44]H. Wesche, W.J. Henzel, W. Shillinglaw, S. Li, Z. Cao, MyD88: an adapter that recruits IRAK to the IL-1 receptor complex, Immunity. 7 (1997) 837-847. [45]R. Tolando, A. Jovanovic, R. Brigelius-Flohe, F. Ursini, M. Maiorino, Reactive oxygen species and proinflammatory cytokine signaling in endothelial cells: effect of selenium supplementation, Free Radic. Biol. Med. 28 (2000) 979-986. [46]G. Bonizzi, J. Piette, S. Schoonbroodt, R. Greimers, L. Havard, M.P. Merville, V. Bours, Reactive oxygen intermediate-dependent NF-kappaB activation by interleukin-1 beta requires 5-lipoxygenase or NADPH oxidase activity, Mol. Cell. Biol. 19 (1999) 1950-1960. [47]K. Hirota, M. Murata, Y. Sachi, H. Nakamura, J. Takeuchi, K. Mori, J. Yodoi, Distinct roles of thioredoxin in the cytoplasm and in the nucleus. A two-step mechanism of redox regulation of transcription factor NFkappaB, J. Biol. Chem. 274 (1999) 27891-22787. [48]E. Pineda-Molina, P. Klatt, J. Vazquez, A. Marina, M. Garcia de Lacoba, D. Perez-Sala, S. Lamas, Glutathionylation of the p50 subunit of NF-kappaB: a mechanism for redox-induced inhibition of DNA binding, Biochemistry. 40 (2001) 14134-14142. [49]M.B. Hampton and S. Orrenius, Redox regulation of apoptotic cell death, BioFactors. 8 (1998) 1-5.

106

R. Brigelius-Flohe and L. Flohe / Glutathione Peroxidases in Signaling

[50]V. Goossens, K. De Vos, D. Vercammen, M. Steemans, K. Vancompemolle, W. Piers, P. Vandenabeete. J. Grooten, Redox regulation of TNF signaling, BioFactors. 10 (1999) 145-156. [51]D.M. Hockenbery, Z.N. Ottvai, X.M. Yin. C.L. Milliman, S.J. Korsmeyer, Bd-2 functions in an antioxidant pathway to prevent apoptosis, Cell. 75 (1993) 241-251. [52]P.A. Sandstrom, J. Murray, T.M. Folks, A.M. Diamond, Antioxidant defenses influence HIV-1 repHcation and associated cytopathic effects, Free Radic. Biol. Med. 24 (1998) 1485-1491. [53]K. Nomura, H. Imai, T. Koumura, M. Arai, Y. Nakagawa, Mitochondrial phospholipid hydroperoxide glutathione peroxidase suppresses apoptosis mediated by a mitochondria! death pathway, J. Biol. Chem. 274 (1999) 29294-29302. [54]M. Brielmeier, J.M. Bechet, S. Suppmann. M. Conrad, G. Laux, G.W. Bomkamm, Cloning of phospholipid hydroperoxide glutathione peroxidase (PHGPx) as an anti-apoptotic and growth promoting gene of Burkitt lymphoma cells, BioFactors. 14 (2001) 179-190. [55]S. Florian, K. Wingler, K. Schmehl, G. Jacobasch, O.J. Kreuzer, W. Meyerhof, R. Brigelius-Flohe, Cellular and subcellular localization of gastrointestinal glutathione peroxidase in normal and malignant human intestinal tissue, Free Rad. Res. 35 (2001) 655-663. [56]R.S. Esworthy, R. Aranda, M.G. Martin, J.H. Doroshow, S.W. Binder, F.F. Chu, Mice with combined disruption of Gpx1 and Gpx2 genes have colitis, Am. J. Physiol. Gastrointest. Liver Physio). 281 (2001) G848-G855. [57]R. Brigelius-Flohe, C. Muller, J. Menard, S. Florian, K. Schmehl, K. Wingler, Functions of GI-GPx: lessons from selenium-dependent expression and intracellular localization, BioFactors. 14 (2001) 101-106. [58]A. Roveri, F. Ursini, L. Flohe, M. Maiorino, PHGPx and spermatogenesis, BioFactors. 14 (2001) 213-222. [59]H. Pfeifer, M. Conrad, D. Roethlein, A. Kyriakopoutos, M. Brielmeier, G.W. Bomkamm, D. Behne, Identification of a specific sperm nuclei selenoenzyme necessary for protamine thiol cross-linking during sperm maturation, FASEB J. 15 (2001) 1236-1238. [60JE. Schmid, A. Hotz-Wagenblatt, V. Hacj, W. Droge, Phosphorylation of the insulin receptor kinase by phosphocreatine in combination with hydrogen peroxide: the structural basis of redox priming, FASEB J. 13 (1999)1491-1500. [61]J.M. Denu and J.E. Dixon, Protein tyrosine phosphatases: mechanisms of catalysis and regulation, Curr. Opin. Chem. Biol. 2 (1998) 633-641. [62]R. Bogumil, D. Namgaladze, D. Schaarschmktt, T. Schmachtel, S. Hellstem, R. Mutzel, V. UHrich, Inactivation of calcineurin by hydrogen peroxide and phenylarsine oxide. Evidence for a dithtol-disulfide equilibrium and implications for redox regulation, Eur. J. Biochem. 267 (2000) 1407-1415. [63]P. Klatt and S. Lamas, Regulation of protein function by S-glutathiolation in response to oxidative and nitrosative stress, Eur. J. Biochem. 267 (2000) 4928-4944. [64]S. Pontremoli, S. Tranielto, M. Enser, S. Shapiro, B.L Horecker, Regulation of fructose diphospnatase activity by disuffide exchange, Proc. Nat). Acad. Sci. U S A. 58 (1967) 286-293. [65]H.Z. Chae, T.B. Uhm, S.G. Rhee, Dimerization of thk>l-specific antioxidant and the essential role of cysteine 47, Proc. Natl. Acad. Sci. U S A. 91 (1994) 7022-7026. [66]H.R. Ellis and L.B. Poole, Roles for the two cysteine residues of AhpC in catalysis of peroxide reduction by alky) hydroperoxide reductase from Salmonella typhimurium, Biochemistry. 36 (1997) 13349-13356. [67]B. Hofmann, H.J. Hecht, L. Ftoh6, Peroxiredoxins, Biol. Chem. 383 (2002) 347-364. [68]M. Saitoh, H. Nishitoh, M. Fujii, K. Takeda, K. Tobiume, Y. Sawada, M. Kawabata, K. Miyazono, H. Ichijo, Mammalian thioredoxin is a direct inhibitor of apoptosis signal-regulating kinase (ASK) 1, EMBO J. 17 (1998)2596-2606. [69]H. Liu, H. Nishitoh, H. Ichijo, J.M. Kyriakis, Activation of apoptosis signal-regulating kinase 1 (ASK1) by tumor necrosis factor receptor-associated factor 2 requires prior dissociation of the ASK1 inhibitor thioredoxin, Mol. Cell. Biol. 20 (2000) 2198-2208.

Thiol Metabolism and Redox Regulation of Cellular Functions A. Pompella et al. (Eds.) IOS Press, 2002

107

Multidrug Resistance-Associated Proteins: Export Pumps for Conjugates with Glutathione, Glucuronate or Sulfate Laszlo HOMOLYA1, Andras VARADI2, and Balazs SARKADI1'3 ' Membrane Research Group, Hungarian Academy of Sciences, Budapest, 2 Institute of Enzymology, Hungarian Academy of Sciences, Budapest, and 3 Institute of Haematology, National Medical Center, Budapest, Hungary 1. Superfamily of ABC Transporters The ATP-Binding Cassette (ABC) transporters form one of the largest protein families; the members of the superfamily can be found in most living organisms from bacteria to man. These proteins are defined by the presence of an ABC unit (or NBD; Nucleotide-Binding Domain), which harbors two short ATP-binding motifs (Walker A and Walker B), and another conserved sequence, the so-called ABC signature motif, which is a hallmark for all ABC transporters. In addition to the sequence homology, these proteins also share distinguishing structural characteristics. They are built from cytosolic nucleotide-binding domains and large transmembrane domains, in most cases composed of 6 transmembrane helices. To our recent knowledge, a functioning ABC transporter consists of at least two transmembrane and two nucleotide-binding units. Some ABC transporters consist of one polypeptide chain (Pgp/MDRl), others form dimers from half transporters (TAP1/TAP2) or are composed from smaller associated subunits [for more information see: ref 1 and 2]. The three-dimensional structure of the mammalian ABC transporters is currently unknown. A low-resolution structure of the MDR1 [3] indicates that the protein is embedded into the membrane as a cylinder with a large central pore and with an opening to the lipid phase. Recent structural studies on a bacterial ABC-transporter [4] first demonstrated the presence of transmembrane helices. After a long period of confusion in naming and numbering of the newly identified ABC transporters, a consistent nomenclature based on sequence homology has been introduced. The approximately 50 human ABC proteins were classified into seven subfamilies, ranging from ABC A to ABCG [see: http://nutrigene.4t.com/humanabc.htm]. Figure 1 depicts the phylogenetic tree of the ABC transporter superfamily. The subfamily ABC A includes ABCA1, the protein that has been proposed as a key component of the reverse cholesterol transport. One of the first identified human ABC transporters, the Pglycoprotein (Pgp or MDR1) belongs to the ABCB subfamily. The Pgp-mediated drug extrusion is still the most widely observed mechanism in clinical multidrug resistance (see below). The ABCC subfamily includes numerous multidrug resistance-associated proteins (MRPs), ABC transporters, which play key role in the cellular elimination of endogenous and xenobiotic lipophilic compounds. In addition, the cystic fibrosis transmembrane conductance regulator (CFTR, or ABCC7), the mutation of which is responsible for the frequent, fatal, inherited disease, the cystic fibrosis, also belongs to this subfamily.

108

L. Homolya el al. / Multidrug Resistance-Associated Proteins

Members of the ABC transporters superfamily are associated with a broad spectrum of physiological functions including detoxification (MDR1, MRP1), defense against xenobiotics and oxidative stress (MRPs), absorption and secretion processes (MDRs,

^ABCC ABCB

C12

C6 (MRP6) C2(MRP2) C1 (MRP1) C3(MRP3) C8(SUR1)

ABCD ABCA

Rg. 1. PhytogeneUc tree of the human ATP-Blndlng Cassette (ABC) transporter proteins. On the basis of sequence homotogy, ABC proteins are classified into seven subfamilies, among which ABCA, ABCB, ABCC, ABCD and ABCG contain the transporter proteins.

MRPs), lipid metabolism (ABCA1, MDR3), antigen presentation (TAP1/TAP2), cell to cell communication (STE6), etc. Most of the ABC proteins are active transporters, mediate uphill transport of substances across the plasma membrane into the extracellular space, or across internal membranes into cellular compartments. This transport activity is driven by the energy of ATP hydrolysis accomplished by the nucleotide-binding domains. Despite the relatively high sequence homology and structural similarity, some ABC transporters possess very different characteristics in terms of their function. In contrast to other ABC transporters, the sulfonylurea receptors (SUR1 and SUR2) modulate the permeability of specific potassium channels. Similarly, CFTR is not an active transport pump, but it forms

L. Homol\a et al. /Multidrug Resistance-Associated Proteins

109

a chloride channel. According to our current understanding, the major function of CFTR is not its chloride channel activity but it plays a regulatory role in controlling the function of other membrane transport proteins [5].

2. ABC Transporters in Multidrug Resistance The most intriguing members of ABC transporters are the so-called multidrug resistance (MDR) proteins, which are associated with the clinical multidrug resistance. These transporters exhibit an unusual broad substrate specificity to a series of hydrophobic compounds used in chemotherapy of cancer. The generally accepted basic mechanism of multidrug resistance is that the MDR proteins actively expel the cytotoxic agents from the cells, maintaining the drug level below a cell-killing threshold. Unlike other, selective (classical) transport proteins, multidrug transporters recognize and handle a wide range of substrates. This wide substrate specificity explains the cross-resistance to several chemically unrelated compounds, the characteristic feature found in the multidrug resistance phenotype [6-8]. Up to now, there are three ABC transporters, which have definitely been demonstrated to participate in the multidrug resistance of tumors: the Pglycoprotein (MDR1, ABCB1), the multidrug resistance-associated protein 1 (MRP1, ABCC1), and the recently identified, mitoxantrone resistance protein (MXR/BCRP, ABCG2) [9-10]. In addition, some other human ABC proteins are able to transport a wide variety of therapeutic agents, they may participate in selected cases of multidrug resistance. These include MDR3 (ABCB4) and sister Pgp or BSEP (ABCB11), and several other members of the MRP subfamily (e.g. MRP2 or ABC2). These transporters are the prime suspects in unexplained cases of multidrug resistance, although proof of their contribution to clinical drug resistance is still missing [for review see 1]. 3. Overview of the Members of MRP Family ABC transporters, belonging to the multidrug resistance-associated protein (MRP) family, have been identified in various species including man, nematodes, yeast, and plants. The human MRP family consists of nine members (ABCC1-6, ABCC10-12). The deduced amino acid sequence lengths range from 1325 acids for MRP4 to 1545 amino acid for MRP2. As a hallmark for the ABCC subfamily, MRP family members contains a 13 amino acid long 'deletion' between the Walker A and Walker B motifs of the N-terminal nucleotide binding domain, in comparison to the NBDs of most other ABC transporters. Although the drug-profiles of multidrug resistance caused by MRP1 and Pglycoprotein (MDR1) seem to be similar, suggesting comparable substrate specificity, these two transport proteins share only 15 % amino acid identity. In addition to the typical, MDR-like core structure consisting of two six-spanner transmembrane units and two cytosolic nucleotide-binding domains, MRP1 has an extra N-terminal transmembrane domain containing 5 membrane-spanning helices (TMDo). The core region and TMDo domain are linked by a shorter cytoplasmic loop (Lo) [11]. It has been shown that TMDo part of MRP 1 is not required for transport activity, but the LO linker region is essential for the function of the protein [12]. With regard to membrane topology, MRP2, MRP3, MRP6 and MRP7 are predicted to be similar to MRP1. In contrast, MRP4 and MRP5 are smaller, they appear to lack the TMDo, and the identity of these proteins with MRP1 is below 40 %. Nevertheless, MRP4 and MRP5 are more homologous to the other MRPs than to MDR1 or

L. Homolya eial. /Multidrug Resistance-Associated Proteins

110

other classes of ABC-transporters. MRP4 and MRP5 possess a long cytoplasmlc Nterminus that is homologous to the LO linker region of MRP 1 (see Figure 2).

Fig. 2. Comparison of the two-dimensional membrane topology models for Pglycoprotein (Pgp) and various multldrug resistance-associated proteins (MRPs).

In

addition to the Pgp-like core structure, most MRPs have an extra N-terminal 5-spanner transmembrane domain (TMDO) linked by a cytoplasrrac loop (LO). TMDO is absent in MRP4 and MRPS.

Four members of the family: MRP1, MRPS, MRP7 and MRP8 are expressed ubiquitously, although MRP1 expression is elevated in the lung and the testes, and reduced in the liver, whereas a higher expression of MRPS can be found in the liver. MRP2 (cMOAT) is predominantly expressed in the canalicular membrane of hepatocytes, and the apical membrane of kidney proximal tubule epithelia. MRP3 can be found in the kidney

L. Homolya et al. /Multidrug Resistance-Associated Proteins

111

and the intestine, MRP4 is expressed in several tissues including prostate, lung, muscle, pancreas, testis, ovary, bladder, and gall bladder, whereas MRP6 expression can be predominantly seen in hepatocytes and the kidney. In contrast to Pgp, in polarized cells MRP1 is localized solely in the basoiateral membrane domain. It has also been demonstrated that MRP3 and MRP6 exhibit lateral expression. However, the expression of MRP2 is limited to the apical membrane of polarized cells.

Table 1. Tissue distribution and cellular localization of the human multidrug resistanceassociated proteins (MRPs) tissue distribution

cellular localization in polarized cells

MRP1

ubiquitous

basoiateral membrane

MRP2

liver, kidney, and gut

apical membrane basoiateral membrane

(high in lung and testis, low in liver)

MRPS

liver, adrenals, pancreas, kidney, and gut

MRP4

prostate, lung, muscle, pancreas, testis,

MRPS

ubiquitous

MRP6

liver and kidney

MRP7

ubiquitous

MRPS

ubiquitous

ovary, bladder, and gall bladder

basoiateral membrane

4. MRP Family of ATP-Dependent Conjugate Export Pumps MRP1, the first member of the MRP family, has been demonstrated to function as a pump for cytostatic agents (e.g. vincristine, vinblastine, doxorubicin, daunorubicine) with a substrate specificity similar to that seen in the case of P-glycoprotein. However, further functional characterization of the multidrug resistance transporters revealed that the preferred substrates for MRP1 are organic anions, including drugs conjugated to glutathione (GSH), glucoronate, or sulfate, whereas P-glycoprotein favors uncharged or positively charged, hydrophobic compounds. Subsequent studies showed that MRP1 is one of the mysterious glutathione S-conjugate pumps transporting endogenous toxic compounds and xenobiotic agents conjugated to GSH out of the cells [13-14]. Typical highaffinity substrates include leukotriene C* (LTC4), bisglucuronosyl bilirubin and 17|3glucuronosyl estradiol. Several other members of the MRP family that have functionally been characterized so far share the property of ATP-dependent export pumps for anionic conjugates and lipophilic anions. Human ABC transporters for which the conjugate transport function has directly been shown include MRP1, MRP2, MRP3, MRPS, and most recently MRP6 [15-18]. Function of MRP-related proteins has been intensively studied in several other species. The non-human transporters for those the ATP-dependent conjugate export pump activity has unambiguously been shown include the murine MRP1 and MRP5, the rat and rabbit MRP2, the yeast cadmium factor 1 (YCF1), and more recently several MRP orthologs in Arabidopsis [18]. Studies on Caenorhabditis elegans provided indirect evidence for that four MRP-related transporters mediate ATP-dependent glutathione Sconjugate transport in this soil nematode [19].

112

L. Homolya et al. /Multidrug Resistance-Associated Proteins

5. Substrate Specificity of Conjugate Export Pumps belonging to the MRP Family The substrate specificity of multidrug resistance-associated proteins has been intensively studied by using inside-out membrane vesicles from cells expressing the recombinant protein at a high level. The detailed functional characterization of the human MRP1, MRP2 and MRP3 revealed that these transporters possess overlapping substrate specificity, but with marked differences in transport kinetics and affinities for various substrates [for review see refs. 16 and 20]. According to the V^/Km ratios, the rank order of the most important substrates of the human MRP1 is as follows: LTC4 > LTD4 > S-(2,4,dinitrophenyl) glutathione > 17(J-glucuronyosyl estradiol > monoglucuronosyl bilirubin > 3ct-sulfatolithocholyl taurine > glutathione disulfide (GSSG). This spectrum of substrates is very similar to that can be found in the case of MRP2 (cMOAT), the ATP-dependent hepatocyte canalicular conjugate export pump. However, MRP1 exhibits a 10-fold higher affinity to leukotriene C4 (Km value 0.1 jxM), and a 10-fold higher affinity to 17(5glucuronyosyl estradiol (Km value 1.5 jxM) in comparison to MRP2, whereas bilirubin glucuronides are preferably transported by the latter. Since a similar range of conjugates is transported by MRP1 and MRP2, it was expected that MRP2 would also be able to mediate the transport of cytostatic agents that are MRP1 substrates. Studies on transfected cells overexpressing MRP2 provided evidence for the ability of this transporter to expel several anti-cancer drugs including methotrexate, vinblastine, etoposide and cisplatine. However, no connection between the expression of MRP2 and clinical multidrug resistance has ever been demonstrated. Despite the overlapping substrate specificity of MRP1 and MRP2, the differences in transport kinetic properties, tissue distribution and cellular localization confer distinctive functions for them. Studies have shown that MRP3, which exhibits a similar tissue distribution as MRP2, but differs in cellular localization (i.e. MRP3 resides in the basolateral membrane of polarized cells) also functions as an ATP-dependent conjugate export pump [21]. In contrast to MRP1 and MRP2, this transporter prefers glucuronate conjugates over glutathione conjugates, which are relatively poor substrates. However, the Kn, value for 17p-glucuronyosyl estradiol was found to be around 70 jiM that is almost 10-fold higher than in the case of MRP2. This observation suggests that MRP3 serves as a basolateral overflow system in hepatocytes for the elimination of toxic agents when the MRP2mediated canalicular secretion is impaired. A detailed analysis using recombinant human MRP3 in membrane vesicles is in progress. The results obtained from this study will make possible a more specific comparison between the substrate specificity of different MRPs. The human MRP4, MRPS and MRP6 have only partially been characterized. An interesting finding revealed, however, that MRP4 can function as an efflux pump for several nucleoside analogues, including drugs that are used against the human immunodeficiency virus such as 9-(2-phosphonylmethoxyethyl) adenine (PMEA) and azidothymidine monophosphate (AZTMP). 9-(2-phosphonylmethoxyethyl) guanine (PMEG), a compound with some neoplastic activity has also been proposed as substrate for this transport protein [22]. Since MRP4-mediated transport of conjugates of GSH, glucuronide or sulfate has not yet been demonstrated, it is possible that MRP4 is specific to phosphate conjugates. Similar to MRP4, MRPS also appears to be a nucleoside analogue pump, conferring drug resistance to PMEA and thiopurines such as 6-mercaptopurine and thioguanine, drugs that are used in the treatment of acute lymphoblastic leukemia (ALL) and acute myeloid leukemia (AML) [17]. However, in contrast to MRP4, glutathione conjugate transport mediated by MRPS has been demonstrated. There has been a controversy in the literature over whether MRPS can confer drug-resistance to heavy metals.

L. Homolya et al. /Multidrug Resistance-Associated Proteins

13

Very recently, functional studies on the MRP6, the mutation, of which causes pseudoxanthoma elasticum (PXE), a rare heritable disorder resulting in the calcification of elastic fibers, has been published. It has been demonstrated that MRP6 can also mediate the transport of glutathione conjugates, i.e. LTC4, and N-ethylmaleimide S-glutathione [18].

6. Role of MRP Isoforms in Detoxification In addition to the contribution of MRP1 (and possibly MRP2) to the clinical multidrug resistance, members of the MRP family have an important physiological role in detoxification. Elimination of endogenous and exogenous lipophilic toxic substances takes place by a multi-step process. The sequence of detoxification comprises the cellular uptake of these compounds (phase 0), followed by an oxidation step (phase 1), conjugation with an anionic group, i.e. glutathione, glucuronate or sulfate (phase 2), transcellular transport (phaseS), and ATP-dependent cellular extrusion of the conjugates (phase 4) [23]. It has been demonstrated that numerous transport proteins belonging to the MRP-family are able to mediate the sequestration of S-conjugates, and contribute to the terminal excretion of lipophilic toxic compounds.

Fig. 3. Detoxification of endogenous substances and xenoblotics. Conjugates with glutathione, glucuronate, or sulfate are expelled by ATP-dependent transport mediated by members of the MRP family.

The importance of MRP isoforms in the process of detoxification becomes apparent in the mild liver disease, Dubin-Johnson syndrome, an autosomal recessively inherited disorder characterized by conjugated hyperbilirubinemia and pigment deposition in the liver. Several mutations in the MRP2 gene have been shown to associate with the absence of the MRP2 protein in the canalicular membrane of the hepatocytes. The lack of MRP2 results in deficient transport of monoglucuronosyl and bisglucuronosyl bilirubin, as well as other anionic conjugates from hepatocytes into bile [24]. The vital importance of MRPs in detoxification and cellular homeostasis is also exemplified by studies in two rat strains, which have been considered as animal models of the human Dubin-Johnson syndrome, the GY/TR- mutant and the Eisai hyperbilirubinermc rat (EHBR). These mutant animals, which lack MRP2 in the hepatocyte canalicular membrane, are unable to sequestrate endogenous metabolites into bile [25-28]. This dysfunction results in the cellular accumulation of the

114

L Homolya et al. / Multidrug Resistance-Associated Proteins

conjugates in the hepatocytes, and a subsequent overflow across the basolateral membrane into the blood. The latter transport step is most probably mediated by MRP3 [28]. Finally, these animals eliminate the LTCU metabolites via renal excretion. Since several members of the MRP family are redundantly present in most cells, in addition to the specialized role of MRP2 in the hepatic function, these proteins seem to prevent the intracellular accumulation of conjugates, contributing to the defense mechanism against lipophilic toxic compounds. Several members of the MRP family may participate in the detoxification of heavy metals, e.g. arsenite and cadmium. The proposed mechanism for the cellular removal of these heavy metals also involves glutathione. For instance, arsenite (HsAsOs) can form a complex with three glutathione molecules, and presumably the glutathione complex of arsenite [As(SG)3] is sequestrated by the transporters. The human MRP1 and possibly MRP5 are able to transport arsenite by this mechanism [29-30], whereas the yeast cadmium factor 1 (YCF1), an MRP ortholog in yeast, confers resistance to cadmium [31]. MRPrelated transporters of the soil nematode, Caenorhabditis elegans contribute to the detoxification of cadmium and arsenite [20]. In summary, the ATP-dependent MRPmediated export of conjugates represents an indispensable terminal step in detoxification.

7. Role of MRP Family Members in Cellular GSH Release Overexpression of MRP1 or MRP2 in cells confers resistance to Vinca alkaloids and anthracyclines. These molecules are weak organic bases, and are not known to be conjugated to acidic ligands in human cells. However, the MRP 1-mediated drug resistance was proven to be dependent on cellular GSH level. In addition, vesicular transport measurements have demonstrated that transport of vincristine, vinblastine, and daunorubicine requires the presence of reduced GSH. The most plausible explanation for these findings is that GSH serves as a co-substrate, and the drugs are co-transported with reduced glutathione [11]. This hypothesis is further supported by the observation that elevated MRP1 expression in tumors is often accompanied by increased expression of yglutamylcysteine syntetase [32]. Similar mechanism was proposed for the MRP-dependent transport of aflatoxinBl [33]. ATP-dependent low affinity transport of GSH has also been demonstrated in yeast. This release of GSH involves the MRP-like transporter, YCF1 [34]. Similarly, the human MRP2 has been proposed to act as a low affinity export pump for the release of GSH across the canalicular membrane into the bile [35]. However, it is not possible to discriminate between the transport of GSH from an MRP2-mediated cotransport of GSH together with an unidentified endogenous substrate. Interestingly, reduced GSH transport mediated by MRP3 has not been observed [36]. It should be noted that additional, non-MRP-like transport proteins involved in cellular GSH release are present in both canalicular and sinusoidal membranes of the hepatocytes. An interesting member of this group is the organic anion transporter OATP1, an antiporter, which resides in the sinusoidal membrane of hepatocytes, and mediates simultanous release of GSH and uptake of hydrophobic Sconjugates such as leukotriene €4 from the blood into cells [37]. 8. Defense against Oxidative Stress The defense machinery that abolishes the elevated level of hydroperoxides under condition of oxidative stress includes GSH peroxidase, producing glutathione disulfide (GSSG), and GSSG reductase, eliminating the increased level of GSSG (see Fig. 4) [38]. Release of GSSG can be observed in several cell types and tissues in response to external addition of

L Homolya et al. / Multidrug Resistance-Associated Proteins

115

hydroperoxides or other conditions causing oxidative stress, indicating that the GSSG reductase pathway may be insufficient. This finding suggest the existence of an overflow mechanism for the elimination of elevated GSSG level [39]. The ability of MRP1 and MRP2 to transport GSSG with low-affinity makes them prime suspect for this overflow system. The GSSG export mediated by MRPs may serve as a mechanism to compensate the oxidative stress. Therefore, these transporters appear to play crucial role in the control of the intracellular GSSG level, when the activity of GSSG reductase becomes rate limiting. Furthermore, this hypothesis is supported by the observation that oxidative stress induces elevation in the MRP1 expression in cultured cells [40].

Fig. 4. Glutathlone dlsulfide (GSSG) transport mediated by members of the MRP-family. MRP1 and MRP2 are low affinity GSSG-transporters. Thus, under condition of oxidative stress, when the activity GSSG reductase becomes rate-limiting, MRPs can expel GSSG from the cells.

9. Outlook Detailed characterization of the transport properties of MRP1 has established the link between membrane transport proteins and the GSH system, previously associated with the clinical resistance to a wide range of anticancer drugs. However, it still remains to be answered whether other members of the MRP family with similar substrate specificity, e.g. MRP2 and MRPS, actually contribute to the clinical multidrug resistance of cancer. The human MRP4, MRPS and MRP6 have only partially been characterized. A detailed analysis, using inside-out membrane vesicles from cells expressing the recombinant proteins at a high level, can provide further information on substrate specificity and transport kinetics of these transporters. The characterization of the new additions of the MRP family, MRP7, MRPS and MRP9 is in progress. Further studies are also needed on non-human MRP orthologs and their role in drug resistance and detoxification of xenobiotics. Numerous controversial issues with regard to transport processes mediated by the members of the MRP family remain to be answered. These include the contribution of MRPs to clinical methotrexate resistance, the cisplatine [(NHs^PtCh] resistance associated with increased MRP2 level [41], the mechanism of MRP2-mediated reduced GSH release from the hepatocytes into bile [37], and the role of MRPS in heavy metal resistance [31]. One of the greatest challenges for future research is elucidating the physiological function of several partially characterized members of the MRP family, such as MRP4, MRPS, and

116

L Homolya et at. / Multidrug Resistance-Associated Proteins

MRP6. Finding the link between the mutations in the genes encoding MRPs and the impaired physiological functions associated with these mutations represents an especially intriguing scientific problem with relevant clinical aspects. Regulation processes of MRP family members are also subject of ongoing research. It has been demonstrated that p53 suppresses the transcription of MRP 1 by diminishing the effect of the transcription activator SP1 [42]. This finding can elucidate the MRP1 upregulation by the loss of p53, as it often occurs in tumors. Further studies are required, however, to identify the regulatory sequences of MRP I gene involved in this process as well as in the upregulation of MRP1 expression in response to oxidative stress [41]. Similarly, the mechanism of transcriptional upregulation of MRPS in the MRP2-deficient rat strain (EHBR) and in patients with Dubin-Johnson syndrome still remains to be disclosed [43, 31]. The protein sorting to the plasma membrane (routing process) represents an important component of the regulation of MRP expression, and has become an intensively studied issue of this research field. The cellular localization of various MRP isoforms greatly influences their function. This is exemplified by the intracellular retention of MRP2 observed in certain cases of Dubin-Johnson syndrome and by the endocytic retrieveal of MRP2 in endotoxin-induced cholestasis [25, 44]. Since it has been postulated the regulatory function of CFTR involves protein protein interactions [45-46], the potential involvement of such interactions in the function and posttranslational regulation of MRP family members is another intriguing issue. In conclusion, despite the fact that considerable knowledge on the MRPs has accumulated since 1992, when MRP1 was identified, the growing number of the MRP family members and non-human MRP orthologs, as well as the complexity of the transport mediated by these proteins represent a great challenge for future research.

References [I]

I. Klein, B. Sarkadi, and A. Varadi, An inventory of the human ABC proteins, Biochim. Biophys. Ada 1461 (1999)237-62. [2] M. Dean, Y. Hamon, and G. Chimini, The human ATP-binding cassette (ABC) transporter superfamHy, J LJpid Res. 42 (2001) 1007-17. [3] M.F. Rosenberg MF, R. Callaghan, R.C. Ford, and C.F. Higgins, Structure of the muKidrug resistance Pglycoprotein to 2.5 nm resolution determined by electron microscopy and image analysis, J. Biol. Chem., 272(1997)10685-10694. [4] G. Chang and C.B. Roth, Structure of MsbA from E. coli: a homotog of the muttidrug resistance ATP binding cassette (ABC) transporters, Science, 293 (2001) 1793-1800. [5] K. Kunzelmann and R. Schreiber, CFTR, a regulator of channels. J. Membr. Biol. 168 (1999) 1-8. [6] C.F. Higgins, ABC transporters: from microorganisms to man, Ann.Rev.Cell Biol., 8 (1992) 67-113. [7] M.M. Gottesman and I. Pastan, Biochemistry of muttidrug resistance mediated by the muttidrug transporter, Annu.Rev.Biochem., 62 (1993) 385-427. [8] M.M. Gottesman., T. Fojo T, and S.E. Bates, Muttidrug resistance in cancer: role of ATP-dependent transporters, Nat. Rev. Cancer, 2 (2002) 48-58. [9] S.E. Bates, R. Robey, K. Miyake, K. Rao, D.D. Ross, T. Lrtman, The role of half-transporters in muttidrug resistance, J. Bioenerg. Biomembr. 33 (2001) 503-11. [10] T. Litman, T.E. Drutey, W.D. Stein, S.E. Bates, From MDR to MXR: new understanding of muttidrug resistance systems, their properties and clinical significance, Cell. Mol. Life. Sci., 58 (2001) 931-959. [II] P. Borst, R. Evers, M. Kool, J. WijnhokJs, A family of drug transporters: the muttidrug resistanceassociated proteins, J. Natl. Cancer Inst. 92(16) (2000) 1295-302.

L Homolya et al. /Multldrug Resistance-Associated Proteins

117

[12] E. Bakos, R. Evers, G. Calenda, G.E. Tusnady, G. Szakacs, A. Varadi, B. Sarkadi, Characterization of the amino-terminal regions in the human multidrug resistance protein (MRP1), J. Cell Sci. 113 (2000) 4451-61. [13] E. Bakos, T. Hegedus, Zs. Hollo, E. Welker, G.E. Tusnady, GJ. Zaman, M.J. Flens, A. Varadi, B. Sarkadi, Membrane topology and glycosylation of the human multidrug resistance-associated protein, J. Biol. Chem. 271(21) (1996) 12322-6. [14] M. Muller, C. Meijer, G.J. Zaman, P. Borst, R.J. Scheper, N.H. Mulder, E.G. de Vries, P.L. Jansen, Overexpression of the gene encoding the multidrug resistance-associated protein results in increased ATP-dependent glutathione S-conjugate transport, Proc. Natl. Acad. Sci. USA. 91(26) (1994) 13033-7. [15] G. Jedlttschky, I. Leier, U. Buchholz, K. Barnouin, G. Kurz, D. Keppler, Transport of glutathione, glucuronate, and sulfate conjugates by the MRP gene-encoded conjugate export pump, Cancer Res. 56(5) (1996) 988-94. [16] D. Keppler, Export pumps for glutathione S-conjugates, Free Radic. Biol. Med. (9-10) (1999) 985-91. [17] J. Wtjnholds, C.A. Mol, L. van Deemter, M. de Haas, G.L Scheffer, F. Baas, J.H. Beijnen, R.J. Scheper, S. Hatse, E. De Clercq, J. Balzarini, P. Borst, MuKidrug-resistance protein 5 is a muftispecific organic anion transporter able to transport nucleotide analogs, Proc. Natl. Acad. Sci. U S A . 97(13) (2000) 747681. [18] A. Ilias, Z. Urban, T.L Seidl, O. Le Saux, E. Sinko, C.D. Boyd, B. Sarkadi, A. Varadi, Loss of ATPdependent transport activity in pseudoxanthoma elasticum-associated mutants of human ABCC6(MRP6), J. Biol. Chem. (2002) epub ahead of print. [19] P.A. Rea, Z.-S. Li, Y.-P. Lu, Y.M. Drozdowicz, E. Martinoia, From vaculoar GS-X pumps to multispecific ABC-transporters, Annu. Rev. Plant Physiol. Plant Mol. Biol. 49 (1998) 727-760. [20] A. Broeks, B. Gerrard, R. Allikmets, M. Dean, R.H. Plasterk, Homologues of the human multidrug resistance genes MRP and MDR contribute to heavy metal resistance in the soil nematode Caenorhabdrtis elegans, EMBO J. 15(22) (1996) 6132-43. [21] J. Kdnig, AT. Nies, Y. Cui, I. Leier, D. Keppler, Conjugate export pumps of the multidrug resistance protein (MRP) family: localization, substrate specificity, and MRP2-mediated drug resistance, Biochim. Biophys. Acta. 1461(2) (1999) 377-94. [22] T. Hirohashi, H. Suzuki, Y. Sugiyama, Characterization of the transport properties of cloned rat multidrug resistance-associated protein 3 (MRP3), J. Biol. Chem. 274(21) (1999) 15181-5. [23] J.D. Schuetz, M.C. Connelly, D. Sun, S.G. Paibir, P.M. Flynn, R.V. Srinivas, A. Kumar, A. Fridland, MRP4: A previously unidentified factor in resistance to nucleoside-based antiviral drugs, Nat. Med. 5(9) (1999)1048-51. [24] T. Ishikawa, The ATP-dependent glutathione S-conjugate export pump, Trends. Biochem. Sci.. 17(11) (1992)463-8. [25] C.C. Paulusma, M. Kool, P.J. Bosma, G.L. Scheffer, F. ter Borg, R.J. Scheper, G.N. Tytgat, P. Borst, F. Baas, R.P. Oude Elferink, A mutation in the human canalicular multispecific organic anion transporter gene causes the Dubin-Johnson syndrome, Hepatology. 25(6) (1997) 1539-42. [26] C.C. Paulusma, P.J. Bosma, G.J. Zaman, C.T. Bakker, M. Otter, G.L. Scheffer, R.J. Scheper, P. Borst, R.P. Oude Elferink, Congenital jaundice in rats with a mutation in a multidrug resistance-associated protein gene, Science. 271(5252) (1996) 1126-8. [27] M. Buchler, J. Konig, M. Brom, J. Kartenbeck, H. Spring, T. Horie, D. Keppler, cDNA cloning of the hepatocyte canalicular isoform of the multidrug resistance protein, cMrp, reveals a novel conjugate export pump deficient in hyperbilirubinemic mutant rats, J. Biol. Chem. 271(25) (1996) 15091-8. [28] K. Ito, H. Suzuki, T. Hirohashi, K. Kume, T. Shimizu, Y. Sugiyama, Molecular cloning of canalicular multispecific organic anion transporter defective in EHBR, Am. J. Physiol. 272 (1997) G16-22. [29] J. Konig, D. Rost, Y. Cui, D. Keppler, Characterization of the human multidrug resistance protein isoform MRP3 localized to the basolateral hepatocyte membrane, Hepatology. 29(4) (1999) 1156-63. [30] G.J. Zaman, J. Lankelma, O. van Tellingen, J. Beijnen, H. Dekker, C. Paulusma, R.P. Oude Elferink, F. Baas, P. Borst, Role of glutathione in the export of compounds from cells by the multidrug-resistanceassociated protein, Proc. Natl. Acad. Sci. U S A . 92(17) (1995) 7690-4. [31] M.A. McAleer, M.A. Breen, N.L. White, N. Matthews, pABC11 (also known as MOAT-C and MRP5), a member of the ABC family of proteins, has anion transporter activity but does not confer multidrug

118

L Homolya el al. /Multidrug Resistance-Associated Proteins

resistance when overexpressed in human embryonic kidney 293 cells, J. Bid. Chem. 274(33) (1999) 23541-8. [32] Z.S. Li, M. Szczypka, Y.P. Lu. D.J. Thiete, P.A. Rea, The yeast cadmium factor protein (YCF1) is a vacuolar glutathione S-conjugate pump, J. Biol. Chem. 271(11) (1996) 6509-17. [33] M.T. Kuo, S. Sen, W.N. Hrttelman, T.C. Hsu. Chromosomal fragile sites and DNA amplification in drugresistant cells, Biochem. Pharmacol. 56(1) (1998) 7-13. [34] D.W. Loe, R.K. Stewart, T.E. Massey, R.G. Deeley, S.P. Cole, ATP-dependent transport of aflatoxin Bt and its glutathione conjugates by the product of the multidrug resistance protein (MRP) gene, Mol. Pharmacol. 51(6) (1997) 1034-41. [35] J.F. Rebbeor, G.C. Connolly, M.E. Dumont, N. Ballatori, ATP-dependent transport of reduced glutathione on YCF1, the yeast orthotogue of mammalian multidrug resistance associated proteins, J. Biol. Chem. 273(50) (1998) 33449-54. [36] C.C. Paulusma, M.A. van Geer, R. Evers, M. Heijn, R. Ottenhoff. P. Borst, R.P. Oude Elferink, Canalicular multispecific organic anion transporter/muMidrug resistance protein 2 mediates low-affinity transport of reduced glutathione, Biochem. J. 338 (R 2) (1999) 393-401. [37] M. Kool, M. van der Linden, M. de Haas, G.L Scheffer, J.M. de Vree, A.J. Smith, G. Jansen, G.J. Peters. N. Ponne, R.J. Scheper, R.P. Elferink, F. Baas, P. Borst, MRP3, an organic anion transporter abte to transport anti-cancer drugs, Proc. Natl. Acad. Sci. U S A . 96(12) (1999) 6914-9. [38] L. Li, T.K. Lee, P.J. Meier, N. Ballatori, Identification of glutathione as a driving force and teukotriene C4 as a substrate for oatpl, the hepatic sinusoidal organic solute transporter, J. Biol. Chem. 273(26) (1998) 16184-91. [39] H. Sies, Intracellular effect of glutathione conjugates and their transport from cells. In: H Sies and B Ketterer (eds.), Glutathione conjugation: mechanisms and biological significance. Academic Press. London. 1988, pp. 175-192. [40] T.P. Akerboom, M. Bilzer, H. Sies, The relationship of biliary glutathione disuffide efflux and intraceNular glutathione disulfide content in perfused rat liver, J. Biol. Chem. 257(8) (1982) 4248-52. [41] Y. Yamane, M. Furuichi. R. Song, NT. Van, R.T. Muteahy, T. Ishikawa, M.T. Kuo, Expression of multidrug resistance protein/GS-X pump and gamma-glutamylcysteine synthetase genes is regulated by oxidative stress, J. Biol. Chem. 273(47) (1998) 31075-85. [42] T. Ishikawa, C.D. Wright, H. Ishizuka, GS-X pump

is functionally

overexpressed in cis-

diamminedichloroplatinum (ll)-resistant human leukemia HL-60 cells and down-regulated by cell differentiation, J. Biol. Chem. 269(46) (1994) 29085-93. [43] Q. Wang, W.T. Beck, Transcriptional suppression of multidrug resistance-associated protein (MRP) gene expression by wild-type p53, Cancer Res. 58(24) (1998) 5762-9. [44] T. Hirohashi, H. Suzuki, K. Ito. K. Ogawa. K. Kume, T. Shimizu, Y. Sugiyama. Hepatic expression of multidrug resistance-associated protein-like proteins maintained in eisai hyperbilirubinemic rats, Mol. Pharmacol. 53(6) (1998) 1068-75. [45] M. Trauner, M. Arrese, C.J. Soroka, Ananthanarayanan M, Koeppel TA, Schtosser SF, Suchy FJ, Keppter D, Boyer JL, The rat canalicular conjugate export pump (Mrp2) is down-regulated in intrahepatjc and obstructive chotestasis, Gastroenterology. 113(1) (1997) 255-64. [46] R. Schreiber. A. Hopf, M. Mall, R. Greger, K. Kunzelmann. The first-nudeotide binding domain of the cystic-fibrosis transmembrane conductance regulator is important for inhibition of the epithelial Ma* channel, Proc. Natl. Acad. Sci. U S A . 96(9) (1999) 5310-5. [47] D.B. Short, K.W. Trotter, D. Reczek, S.M. Kreda, A. Bretscher, R.C. Boucher, M.J. Stutts, S.L Mikjram, An apical PDZ protein anchors the cystic fibrosis transmembrane conductance regulator to the cytosketeton, J. Biol. Chem. 273(31) (1998) 19797-801.

Thiol Metabolism and Redox Regulation of Cellular Functions A. Pompella et al. (Eds.) IOS Press, 2002

119

Detoxification of Electrophilic Compounds by Glutathione S-transferase Catalysis: Determinants of Individual Response to Chemical Carcinogens and Chemotherapeutic Drugs? Brian F. COLES and Fred F. KADLUBAR National Center for Toxicological Research, Jefferson, Arkansas, USA. E-mail: [email protected] Tel: 870 543 7596 Fax: 870 543 7773

1. Introduction The glutathione S-transferases (GSTs) catalyze the GSH-dependent detoxification of electrophilic xenobiotics and certain products of oxidative stress, and are, thus, regarded as being critical for cellular homeostasis. There are two families of GSTs. The family of soluble GSTs comprises at least 16 genes in humans, grouped into 8 classes. The second family, the microsomal GSTs, are structurally unrelated to the soluble forms and are, thus, not considered further in this discussion; hereafter, "GST" refers specifically to the soluble forms. The genetic organization of the human GSTs is given in Table 1. The GST pi class contains a single gene but the alpha, mu and theta classes form clusters of closely related genes that code for closely related GST subunits. The catalytically active proteins are dimers of subunits from within the same class. The GSTs of the alpha, mu, pi and theta classes, which have been most extensively studied, are known to accept a wide range of electrophilic substrates. These include genotoxic carcinogen metabolites, industrial pollutants and chemicals, chemotherapeutic drugs, halogenated solvents and fatty acid hydroperoxides. Because genotoxic and cytotoxic electrophiles are substrates for these GSTs, they are of particular interest with respect to susceptibility to cancer and the efficiency of chemotherapeutic agents. It is important to note that GST isoenzymes show selectivity of electrophile acceptance (see Table 1 and [1-3, 5] for examples) and that not all electrophiles are substrates for the GSTs. (For reviews see [1-6].) Several allelic polymorphisms occur in the human GST genes. These are known to affect protein expression or to code for proteins with variant catalytic properties (Table 1). The GSTM1 "null" and GSTT1 "null" polymorphisms are deletions of the GSTM1 or GSTT1 genes, and individuals who are homozygous null cannot express the corresponding protein [1,4, 6-7]. A polymorphism in intron 6 ofGSTMS [4, 8] appears to have a minor effect on GSTM3 expression [4, 9]. Polymorphism in the promoter of GSTA1 affects GSTA1 expression in the liver, apparently via the levels of basal

Table 1. Human soluble glutathlone S-transferases: genetic organization, examples of biologically relevant substrates and allellc polymorphism. CLASS Alpha

CHROMOSOMAL LOCATION [4] 6p12

GENES/ BIOLOGICALLY IMPORTANT SUBSTRATES PROTEIN and REACTIONS

POLYMORPHISM

CONSEQUENCES OF POLYMORPHISM

GSTA1

GSTAI'A. GSTArB, 3 linked SNPs in 5'-regulatory region [8].

Determines ratio of GSTA1/GSTA2 in liver; GSTAI'B causes 4x lower GSTA1 expression in liver [10]. Not known. Little affect on known activities [12].

GSTA2 GSTA3

N-acetoxy PhIP [26]. PAH dralepoxides (low cf. GSTP1) [1, 2], phosphoramide mustard [5]; peroxidase towards fatty acid hydroperoxides [1, 3]. Peroxidase towards fatty acid hydroperoxides.

GSTA1 variants: SNPs [12]. GSTA2'A -*D; 3 SNPs resulting in amino acid substitutions and 4 haplotypes [12]. None known

GSTA4

Isomertsation of A5-androstene-3,17-dione to testosterone and A5-pregnene-3,20-dk>ne to progesterone [42]. 4-hydroxy-2-nonenal [4].

None known

Kappa

Not known

GSTK1

Not known

None known

Mu

1p13.3

GSTM1

Aflatoxin B1 epoxide [1]; PAH epoxides and PAH diotepoxides (Lowcf. GSTP1)[1J. Organic isothiocyanates [23]. Alkenab[1,3]

GSTMI'A and GSTMI'B (G519>C (Lys173>Asn) [1,4].

No significant effect on known activities [1,4].

GSTMfO (= GSTM1-null); gene deletion[1,4, 7] GSTMn. gene duplication [1,4].

No protein expression [1,4,7]. High protein expression [1,4].

GSTM2

Aflatoxin B1 epoxide [1].

GSTM3

BCNU [1]

GS7M3*e, 3 bp deletion in intron 6; linkage disequilibrium with GSTMI'A [8].

Potential of higher expression in GS7W8[8].

PAH diol epoxides, acrotein [1,2], 4-OH-cydophosphamide [5], thiotepa [43], DNA-aktehyde adducts [1. 3]. PGE2synthase[1,4].

GSTPI'A -"D; 2 SNPs; A313>G, C341>T (amino add changes lle105>Val, Ala114>Val); and 4 haplotypes [4.11]. (Silent mutation C555>T [4].)

Differences in catalytic efficiency towards PAH diol epoxides and other substrates [111. thiotepa [431.

GSn*0(-GSTT1-null)[7]

No protein expression

GSTM4 GSTM5 11q13

GSTP1

Sigma

4q21-22

GSTS1

Theta

22q11

GSTT1

Zeta

14q24.3

GSTZ1

Butadiene epoxides, ethytone oxide [6]. Activation of methyl chloride, trichtoroethytone, dibromoethane [6]. Benzylic sulphates (e.g., 7-hydroxymethyl,12methylbenzathnicene derivatives) [6]. Dichloroacetate, fluoroacetate [4].

Omega

10q23.5

GST01

Thioltransferase[4]

GSTT2

None known GSTZI 'A ••D; 3 SNPs producing 4 haplotypes [4]

Altered substrate specificity [7].

Details of several aspects of GST biochemistry and references to additional original sources can be found in [1-7] and in references quoted in [11,43]. PAH = polycyclic aromatic hydrocarbon. SNP = single nucleotide polymorphism.

! i

B. F. Coles and F. F. Kadluba r / Detoxification of Electrophilic Compounds

121

expression [10]. Polymorphism in the coding region of GSTP1 results in proteins with variant catalytic properties, notably towards carcinogenic polycyclic aromatic hydrocarbon diol epoxides [1-2, 4] (and see references in [11]). Polymorphism in GSTA2 [12] and GSTZ1 [4] also produce variant proteins but these have been little studied to date. Genetic polymorphisms are readily determined for large study populations using DNA isolated from whole blood or archival histological tissue. Thus, these polymorphisms offer a tool that can be used in epidemiological studies to explore the hypothesis that GSTs play a role in determining individual response to chemical carcinogens and chemotherapeutic drugs. Epidemiological studies do not of course prove any causal relationships. Nevertheless, the associations that have been established between allelic polymorphism of GSTs and susceptibility to cancers are usually interpreted in the light of the polymorphism as a predictor of GST activity and there being a role (or not) for the GST in the etiology of disease. A list of cancers for which an association between risk and GST genotype has been established in at least one epidemiological study is given in Table 2. Examples of more complex apparent interactions are given in Table 3. A feature of these studies is the moderate associations of cancer risk with GST polymorphisms, the variability of result between different study populations and the dependence of the degree of risk on other population characteristics. Does this indicate that, despite their proven catalytic properties and demonstrable anti-carcinogenic properties in experimental animals [1], GSTs are not major factors in the etiology of human cancer? In the summary that follows, several aspects of GST expression are discussed, using examples taken primarily from our own studies. These examples show how populations are more variable as regards the expression and role of GSTs in carcinogenesis (or response to other genotoxic or cytotoxic agents) than would be indicated by genotype alone. "Mechanistic models" for chemical aspects of colorectal carcinogenesis and the efficiency of combination chemotherapy for treatment of breast cancer are given as examples of ways to identify potential risk factors and relevant study populations for epidemiological studies. Table 2. Cancers for which an association with a GST allelic polymorphism has been found In at least one study1 GST M1

Cancer Basal- and squamous-cell carcinomas, bladder, breast, cervix, childhood acute leukemia, colorectal, esophagus, larynx, liver, lung, meningioma, mesothelioma, nasopharyngeal, oral, ovary, pituitary, prostate, stomach.

T1

Astrocytoma,

basal-

and

squamous-cell

carcinomas,

bladder,

breast,

cervix,

childhood

acute

lymphoblastic leukemia, colorectal, esophagus, kidney, liver, lung, meningioma, oligodendrioma, ovarian, prostate2. P1

Breast, bladder, esophageal, kidney, lung, malignant glioma, oral, pharyngeal, prostate, testicular, squamous cell carcinoma.

M^

Basal- and squamous-cell carcinomas, larynx, lung, oral.

A1

Colorectal

'An association may have been observed only in a single study and not confirmed by subsequent studies. Similarly, associations may be limited to certain subgroups of the study populations (see Table 3 for some examples). 2GSTM1-nuil associated with a decrease in risk. Results are taken primarily from reviews [4, 6, 7).

Table 3. Examples of associations between GST polymorphisms and cancers, selected to Illustrate some of the complexity of Interactions1 Cancer Bladder

GST GENE M1-null

Risk alone2

Comments and Apparent Interactions

Not significant to-3.8 fold increased risk [7].

Most consistent of the associations of GST genotype with a cancer susceptibility [7]. Slight increase in smokers; NAT2 slow [7].

M3*A

Not significant

Slight increase in risk in combination with GSTM1-null [7].

T1-null

Increased risk of 3-4 (Egyptian population)

M1-null

Increased risk of 1.75; P1 A/G or P1 GIG cf. P1 A/A (Turkish population) [43]. Not significant to ~2x risk [7]

M3*A

Not significant alone [45].

T1-null

Not significant alone [6]

P1 Lung

Colorectal

Breast

Not smoking dependent in this population; schistosomiasis [6]. 3

Smoking + GSTM1-null increases risk to 3.9 [43]. Most significant for squamous cell carcinoma and non small cell carcinoma; Less so or not for adenocarcinoma [7,45]. Most significant for moderate to heavy smokers [7]. Both GSTT1 null and GSTM1 null increase risk [6,45]. Strong interaction with CYP1A1 and CYP2E1 [6, 7J. Interaction with GSTM1-null, GSTT1-null and GSTP1 [45]. GSTTI-null + GSTM1-null increase risk [6, 45].

J

P1

P1 GIG or A/G cf. P1 A/A or P1 A/G ; increased risk of 1.7-2.4. [45].

Not consistent between studies [4]; increase in risk with combined 'low activity'' GST alleles [45].

M1-null

Not significant to ~2 (see Rets. In [26).

Correlation depends on tumor site; proximal but not distal tumors [6, 7]. Associated with early age of onset [7].

T1-null

Not significant [6].

Earlier age of onset [6]. More pronounced in NAT2 stow individuals [6].

A1*B

Increase risk of -2 for A1*B homozygotes cf. all other genotypes [26].

Risk associated with dietary well-cooked meat [29].

M1

Little effect atone [7, 46].

Most evident association in post-menopausal women; null effect not altered by smoking or dietary anttoxktents [46]. Interaction with GSTT1 and GSTP1 [6, 46].

M3

Little effect atone [46].

T1

Not clearly demonstrated [6].

Accelerated age of onset [7]. Interaction with GSTM1, GSTTM3 and GSTP1 [6, 46].

P1

PVATD cf. P1*BTC no risk (see refs. in [47]). PVC cf. other genotypes; decreased risk of ~2 [47].

Inconsistent between studies ; interaction with GSTT1, GSTM1, GSTM3 [4, 6, 46,47],

1 These examples have been selected to illustrate some of the complexity of apparent interactions of genotype for cancers discussed in the text. The summary should not be regarded as a representative review of the literature. Further details and references to original sources can be found in reviews [4, 6, 7] and in the references quoted in more recent accounts [26, 44-47]. The risks quoted are the odds ratios associated with the specified alleles The nomenclature refers to the base change A313>G.

En

I

B.F. Coles and F.F. Kadlubar / Detoxification of Electrophilic Compounds

123

2. Variability of GST Expression GSTs have been found to be expressed in all organs studied. However, their levels of expression vary widely both between organs and between individuals [1, 13]. In our studies, GSTs have been isolated from cytosols by GSH-agarose affinity chromatography and quantitated (as subunits) by HPLC (Fig. 1) [9]. This method has the advantage that all GSTs of the alpha, mu and pi classes are separated and quantitated simultaneously, including closely related subunits that are not readily distinguished immunochemically. GSTs of the theta and zeta classes are not retained by GSH affinity matrices and our examples are restricted to GSTs of the alpha, mu and pi classes.

a. PANCREAS

P1 A214

35

38

41

45

49

time (min)

Fig. 1. HPLC separation of alpha, mu and pi class GSTs of human pancreas, liver and colon by HPLC. GSTs have been isolated by GSH-agarose affinity chromatography and reverse-phase HPLC from normal human tissues (see [9] for details). Note the separation of GSTMIa from GSTMIb and GSTM3, and GSTA1 from GSTA2; subunits that are not readily distinguished immunochemically. GSTs have been isolated by GSH-agarose affinity chromatography and reverse-phase HPLC from normal human tissues (see [9] for details). Note the separation of GSTMIa and GSTMIb and GSTM3 and GSTA1 and GSTA2 that are not readily distinguished immunochemically.

124

B.F. Coles and F.F. Kadlubar / Detoxification of Electrophilic Compounds

2.1. Tissue Specific GST Expression Organ-specific patterns of GST expression can be strikingly different. For example, GSTA1 and GSTA2 are expressed at high levels in human liver (Fig.lb), and can represent 3% of total cytosolic protein [13, 14]. In contrast, GST expression in colon is particularly low (<0.2 %) [15] and consists primarily of GSTP1 (Fig.lc). This pattern of predominant GSTP1 expression is shown for lung and many other tissues [1, 13]. GSTM1 is expressed at high levels in few tissues, notably liver, testis, brain [13] and bladder [16]. In colon (Fig. Ic) and lung [11, 17], it is a minor component expressed at lower levels than the closely related GSTM3. Tissue-specific expression is important in that it may predispose certain organs to the genotoxic effects of chemical carcinogens because they lack a GST with (selective) activity towards a critical chemical carcinogen [1,2]. Conversely, if a GST is not expressed in an organ then a genetic polymorphism will not be effective in that organ although it may be of relevance elsewhere and have an impact indirectly.

r = 0.68, P <0.001

0.06 c '(D O Q.

CO o ^ o

0.04

r^ o |_ (0

O)

0.02 -

0.00

0.0

0.5

1.0

1.5

2.0

2.5

GSTP1 (ng/mg cytosolic protein)

Fig. 2. Expression of GSTP1 and GSTM3 In human lung. GSTs have been quantitated in 29 samples of normal tissue by HPLC. Note the approximately 7-foW variation in levels of expression between individuals and the degree of conservation of ratios of expression.

B.F. Coles and F.F. Kadlubar /Detoxification of Electrophilic Compounds

125

2.2.1. Interindividual Variation of GST Expression and the Consistency of Organ-specific Patterns of Expression We have examined several tissues in detail to establish the range of "normal" variation of expression. Notable is the conservation of organ-specific patterns of GST subunit expression, despite the variability of absolute values of expression. Expression of GSTP1 and GSTM3 in lung (Fig 2) [11] is an example of what we regard as conserved expression. This is illustrated further by GST expression in the pancreas [9] (Table 4). For GSTP1, GSTA2 and GSTM3, the variability in terms of per cent composition is less, or similar to, that of the absolute values of expression. These three GSTs appear to be the constitutive elements subject to a high degree of effective co-regulation (although not necessarily by the same mechanism [1]). Variation greater than ~7-fold in % composition (e.g., GSTA1 and GSTM1 in this tissue) appears to indicate some form of polymorphism, either genetic (e.g., GSTM1-positive individuals include heterozygotes and homozygotes) or due to other factors such as induction (see below).

Table 4. Range and fold variation of GST subunit expression in samples of normal pancreas. GST subunit

range (pg/mg cyt Prot.) mean fold variation composition (%) mean fold variation

M3

all GST

P1

A1

A2

M1

M2

1.1-6.5

0.18-5.1 1.44 28.3

0.82-10.8

0.04-0.5 0.17

0.02-0.33

0.12-0.91

2.8-20.3

4.62

0.07

0.34

9.24

13.2

12.5

16.5

7.58

7.25

18.7-57.6

0.55-7.98 2.05 14.5

0.28-4.66

1 .24-8.0

1.23

3.86

16.6

6.45

2.78 6.13

15.3-67.3 32.2 4.4

1.36-41.0 15.3 30.14

40.9 3.08

GSTs of the alpha mu and pi classes have been isolated from 43 normal tissue cytosols by GSH-agarose affinity chromatography and quantitated by HPLC [9].

2.2.2. GSTP1 and Lung Cancer The effect of the GST PI genotypes is usually interpreted as being due to the differential activity of the variant allele protein products towards carcinogenic polycyclic aromatic hydrocarbon (PAH) diol epoxides. PAHs are lung carcinogens in experimental animals, are present in cigarette smoke, and smoking is a major risk factor for lung cancer [18]. Consequently, the association of the GSTP1 alleles with risk of lung cancer has been investigated in a number of populations. An increased risk of lung cancer according to GSTP1 genotypes has been found, but results are inconsistent between studies [4] (and see Table 3). The differences in catalytic efficiency between GSTPla and GSTPlb, the protein products of the more common (hence most frequently investigated) alleles, is moderate (e.g., ~7-fold for BPDE [11]). This range of variation can cause deduction of catalytic activity on the basis of genotype to be inaccurate because normal interindividual variation of GSTP1 expression is also at least 7-fold (Fig 2). An individual with low expression of an efficient GSTP1 variant can have lower activity than an individual with higher expression of a lower activity variant. The GSTP1 alleles do not differ in their 5'-regulatory regions (see refs. in [11]), and we found

126

B.F. Coles and F. F. Kadlubar / Detoxification of Electrophilic Compounds

similar levels and ranges of variation of expression in GSTP1*A/*A and GSTP1*B/*B individuals [11]. The inaccuracy of deduction of GST activity on the basis of genotype alone would lead to confounders in epidemiological studies and reduce the observed significance of the genotype on cancer risk. The effects of GSTP1 genotype is further complicated because tobacco smoke contains many genotoxic compounds of varied structure [18], and differential substrate specificity of the GSTP1 protein variants is highly dependent on electrophile structure (see refs.in[ll]).

5J

5

4H

8 O

3-

GSTM1*0/*1or*1/*1 (positive)

Q. O

O O)

!>

GSTM1*0/*0 (null)

1-

10

20

30

40

sample number

Fig. 3. Variation of GSTM1 expression in liver. GSTM1 protein (i.e., GSTMIa or GSTMIb. see Fig. 1) was quantitated for 55 normal liver samples. GSTM1 genotype was determined for the same samples. Note the approximately 50-fold variation in expression.

2.2.3. GSTM1 and Smoking Related Cancers The associations of GSTMl-null with increased risk of the smoking related bladder and lung cancers have proved to be the most consistent relationships between a GST polymorphism and cancer [7]. GSTM1 is expressed at high (but variable) levels in the bladder [16] suggesting that local GSTM1-dependent detoxification can occur in this organ. The GSTMl-null polymorphism is, therefore, highly relevant to the bladder, and this may be one reason why an association between GSTMl-null and increased susceptibility to bladder cancer is frequently observed. However, because GSTM1 is expressed at low levels in lung (c.f., Fig. Ic, [11, 17]), the GSTM1 polymorphism would not be expected to be important in this organ (but see

B.F. Coles and F.F. Kadluhar / Detoxification of Electrophilic Compounds

127

below). GSTMl expression in the liver (Fig. Ib) may be more important in this case, because liver is a major site of carcinogen metabolism and expresses GSTMl at relatively high levels. GSTMl expression in normal liver is highly variable (Fig. 3). Whereas the homozygous GSTMl-null genotype is well defined phenotypically by absence of protein, the GSTMl-positive phenotype is not and shows an -50 fold variation in expression. The GSTMl positive genotypes would be expected to represent a heterogeneous group, again acting as a confounder in epidemiological studies. 2.2.4. GSTM3 in Linkage Disequilibrium with GSTMl The GSTM3*B genotype has been shown to be in linkage disequilibrium with GSTMl*A. That is, there is an over-representation of GSTM3*B individuals (putative higher expression in lung via a modified intronic regulatory element) who are GSTMl*A (i.e., the most common 1.0

GSTM1*A c '0 o

„ Q- 0.6

CO o

C/D

& 0.4 -

GSTM1*0/*B

O)

E

o 0.2 -

0.0

GSTP1 M-g/mg cytosolic protein Fig. 4. Correlation of GSTP1 and GSTM3 expression In pancreas according to GSTM1 type. GSTP1 and GSTM3 have been quantitated in samples of human pancreas by HPLC. GSTM1 type has been assigned on the basis of phenotype as GSTMta (•), null (no protein) and/or GSTMfb (O) (see Fig. 1 for separation). Note the correlation for the data as a whole (r ~ 0.7, P <0.001) and the improved correlation for GSTMIa individuals (r = 0.94, P <0.001). The GSTMIa samples include presumed homozygous GSTM1*A and heterozygous GSTMTA/"0 individuals. The consistency of linkage of GSTM1*A and GSTM3*B genotype is not known in these samples.

GSTMl-positive allele [4, 8]). An example of an improved correlation between GSTM3 and GSTP1 expression according to GSTMIa phenotype is shown for pancreas (Fig. 4). A similar

128

B.F. Coles and F.F. Kadlubar / Detoxification of Electrophilic Compounds

analysis for the lung is not available, but there is evidence that GSTM1-positive individuals do have higher GSTM3 in the lung than GSTMl-null individuals [17]. Thus, the effect of the GSTMl-null genotype appears to be compounded by low GSTM3 expression. As noted above, GSTM3 is expressed at higher levels than GSTM1 in lung (~ 2-fold in our studies [11]). Thus any "effect" of the GSTM1 -positive genotype on susceptibility to lung cancer may be confounded or even likely due to GSTM3 expression. GSTMla, GSTMlb and GSTM3 have similar catalytic properties [1,4]. 2.2.5. GSTA1: Polymorphisms are not Functional in All Tissues Expression of GSTA1 appears to be particularly variable both between organs and between individuals. A comparison of the expression of GSTA1 and GSTA2 in liver and pancreas (organs that both express GSTA1 and GSTA2 at high levels) according to the polymorphism in the 5'-regulatory region of GSTA1 is given in Fig. 5. For the pancreas, the ratio of GSTA1 to GSTA2 is reasonably correlated for the whole data set and the polymorphism does not appear to have much effect. Conversely, for the liver there is no correlation for the data as a whole but

a. LIVER

b. PANCREAS

14

A/A 12

•I 10 o Q_

i- o < o OD S O o en

A/B

8 -

6-

B/B

14 -

2 -

2

4

6

8

GSTA2 ng/mg cytosolic protein

10

0

2

4

6

8

10

GSTA2 ng/mg cytosolic protein

Rg. 5. Expression of GSTA1 and GSTA2 In liver and pancreas according to GSTA1 genotype. GSTA1 and GSTA2 have been quantitated for 55 liver samples and 45 pancreas samples using normal tissue from organ donors. For liver, note the lack of correlation of GSTA1/GSTA2 expression for the whole data set (r = 0.06, P = 0.65) but the significant correlations according to GSTA1 genotype (r = 0.78, P = <0.001 for GSTA1*A; r = 0.55, P = 0.005 for heterozygotes, similar to those shown in Fig 2). For pancreas, note that GSTA1/GSTA2 are poorly correlated for the whole data set (r =0.2, P <0.001) and that analysis by genotype does not markedly improve the correlation. • (AA) = heterozygous GSTA1*A, O (A/B) = heterozygotes, T(BB) = homozygous GSTATB. For genotyping methods see [10].

B.F. Coles and F.F. Kadlubar / Detoxification of Electrophilic Compounds

129

correlations become significant according to GSTAl genotype. The polymorphism also correlates with an approximately 4-fold difference in mean hepatic expression of GSTAl [10]. We interpret the differential effects as a result of the dominance of tissue-specific regulatory elements vs. that of the polymorphic element, but it is not known how these interact. Thus, the polymorphism would be expected to be important for a carcinogen that is detoxified by GSTAl in the liver, but not for detoxification in the pancreas. It is not known if the polymorphism is active in other tissues that express GSTAl at high levels (e.g. kidney, testis and small intestine [13, 15]). 2.2.6. GST Induction By using human cells in culture, it has been shown that GSTs are inducible [19], notably by organic isothiocyanates derived from cruciferous vegetables [20-21]. This is thought to be one way by which diets high in vegetables can reduce the risk of cancers [22]. A combination of GSTM1-positive homozygotes, heterozygotes and GSTM1 induction [21] might be the cause of the variability of hepatic GSTM1 expression shown in Fig 3. The wide variation of GSTAl expression in pancreas (Table 4) might also represent induction as the GSTAl polymorphism referred to above is not active in pancreas. (Note also the three high GSTAl outliers in Fig. 5b). The effects of dietary inducers are not independent of genotype. For example, volunteers given a broccoli diet had significantly increased alpha class GSTs in blood plasma (interpreted as arising from the liver) only when they were GSTMl-null [21]. This is attributed to the action of GSTM1-dependent detoxification of isothiocyanates [23] that destroys their inductive potential. Thus, the protective role of GSTM1-positive genotype in detoxification of chemical carcinogens runs counter to its effect on the response to GST alpha inducers. Therefore, the effect of GSTM1-genotype will be highly dependent on diet and other factors in the etiology of any particular cancer. 3. Mechanistic Model for Chemical Aspects of the Etiology of Colorectal Cancer Most environmental and dietary carcinogens are genotoxic only after activation in vivo. The cytochrome P450s (CYPs) are the most efficient enzymes of carcinogen activation and because these are expressed most highly in the liver, the liver is thought to be the primary site of carcinogen activation. CYP-catalyzed oxidation is not always sufficient for activation, but is frequently followed by conjugation by, e.g., N-acetyltransferases and sulphotransferases (notably for heterocyclic amine carcinogens; see refs. in [24-26]). Our model [24-25] for chemical aspects of colorectal carcinogenesis (CRC) is based on several known risk factors for CRC. The fact that heterocyclic aromatic amines (HAAs) formed during the cooking of meats are colon carcinogens in experimental animals, the known pathway of HAA activation to genotoxic electrophiles, and the association of CRC in humans with high levels of cooked meat in the diet. Sequential carcinogen exposure, CYP1A2catalyzed N-oxidation and a secondary N-acetyltransferase-2 catalyzed O-acetylation are required before any genotoxic effect can become apparent. In the version of the model

130

B. F. C o l a and F.F. Kadlubar / Detoxification of Electrophilic Compounds

flg. 6. Mechanistic modd for chemkal 88pect8 of colomctal cancer (CRC) baaed on metabolbm of the pmdomlnant mat-dedvedcarcinogen PhlP.

presented here (Fig. 6), HAA carcinogens are represented by PhIP. N-acetoxy-PhIP is one of the few amine carcinogens known to be a substrate for any GST [26], and PhIP is the major HAA formed during cooking of meats [27]. If this pathway represents a significant contribution to the etiology of CRC then susceptibility to colorectal cancer should depend (in part) on the efficiency of each step. The studies of Lang et al. [25] and Le Marchand et al., [28] were designed to address this hypothesis. In these studies, each individual was assessed as a “slow” or “rapid ” CYPIA2 andor NAT2 metabolizer by non-invasive phenotyping (based on specific ratios of urinary metabolites of caffeine) or genotyping. In both studies, the risk of developing CRC (or colorectal adenomas) increased with putative high carcinogen exposure (based on a preference for well-done meat) and increased further with the number of more efficient activation steps. These results give support to the hypothesis that the pathway described above represents a significant contribution to the etiology of CRC. This model is of course a simplification of the etiology of CRC. Apart fiom the important events of DNA repair, apoptosis etc. that occur after genotoxic insult, there are other detoxification and activation processes that have not been included in this version of the model, (N.B., detoxification of N-hydroxy-PhIP by glucuronidation [24-261). Nevertheless, it can be seen fiom this scheme that interindividual variation of CYPIA2 activity or NAT2 activity would be of little consequence for susceptibility to CRC without carcinogen exposure. Similarly variation of GST activity can only be of consequence when activated carcinogen is present, both for this model and for chemical carcinogenesis in general. Populations will be highly heterogeneous for exposure relevant to GST activity. The polymorphism in GSTAl would appear to be highly relevant to CRC because GSTAl is the only GST that efficiently detoxifies N-acetoxy-PhIP [26]. N-acetoxy PhIP is stable enough that it can be formed in the liver [24] where the GSTAl polymorphism is known to be active [lo], and can be transported to the colon, the target organ, where GSTAl is expressed at critically low levels (Fig Ic) [15]. For our study population as a whole (see Fig. 7 for details), the GSTAl*B allele (which confers lower hepatic GST expression in liver) was associated with an -2-fold increase in susceptibility to CRC [26]. However, analysis according to preference for well done meat, showed that this association was evident only for the higher

to preference for well done meat, showed that this association was evident only for the higher dietary meat category, for which risk increased to -3 (Fig. 7) [29]. That is, the effect is seen only in the population with putative higher PhIP exposure. GSTMl genotype is also relevant to this model because it appears to be a modifier of GST alpha induction (see above). The possibility exists that GSTMl-null is anti-carcinogenic in individuals with high cruciferous vegetable intake via GST alpha induction in colon, but procarcinogenic in other populations. There is evidence that this is the case for susceptibility to colorectal adenomas (the CRC precursor lesion) where an effect of diet was reported primarily for GSTM1-null individuals [301. The complex interaction and variability of these factors between individuals may explain some of the inconsistency of result between studies of GST genotype and susceptibility to CRC.

Fig. 7. Odds ratlo for risk of developing colorsctal canmr according to GSTAl genotype and red meat consumption. The population consisted of 114 cases of colorectal cancer, matched for age, sex, geographic region, race, current smoking, education and total red meat intake with 259 controls. Cooking preferences and intake of six red meats was assessed by an interview. For the population as a whole, the odds ratios (95 % confidence interval) for risk of disease associated with homozygous GSTAl'5 (BB) cf. homozygous GSTAl *A (AA) plus heterozygotes (AB) is 2.0 (C.I. 1.O3.7).For the lower meat category and the same analysis, the odds ratio = 1.4 (C.I. 0.6-3.0), and for the higher meat category 3.3 ((3.1. 1.2-8.9); the low meat and GSTAlA" plus heterozygotes being used as reference.

4. GSTS and Allcylating Chemotherapeutic Agents

In the case of alkylating chemotherapeutic agents the role of GSTs may be better defined, because all individuals are exposed to acute dose(s) of the chemotherapeutic agent which is

132

B.F. Coles and F.F. Kadlubar / Detoxification of Electrophilic Compounds

either direct acting or requires a single activation step, thus avoiding much of the variability of relevant exposure. The effect of GSTs in chemotherapy will be the reverse of that in cancer susceptibility, i.e., detoxification of the chemotherapeutic agent would decrease chemotherapeutic efficacy. We have examined survival of breast cancer following adjuvant chemotherapy according to genotypes of GSTA1, GSTP1, GSTM1 and GSTT1. All individuals in the study received cyclophosphamide (CP) in combination with adriamycin, methotrexate or 5fluorouracil. Polymorphism in GSTA1 and GSTP1 would appear to be highly relevant to the study because GSTA1 is known to catalyze the detoxification of the therapeutic metabolite of CP, phosphoramide mustard (PM), and GSTP1 catalyzes the detoxification of 4-hydroxy-CP, the precursor of PM [5]. In this study, GSTA1*B/*B was correlated with greater survival than GSTA1*A/*B or GSTA1*A/*A [31]. Similarly, the GSTP1 homozygous "Val" genotype was found to be correlated with greater survival than GSTP1 "He" [32]. GST alpha expression in breast tumor did not correlate with survival (unpublished results of authors of [30]) but GSTP1 expression did [33]. Interestingly, GSTMl-null and GSTTl-null individuals also showed greater survival [34] even though a direct role of these GSTs in metabolism of CP or other components of the chemotherapy is not known.

Table 5. Associations between GST polymorphisms and survival after chemotherapy for breast cancer GST gene/genotype

Number /genotype

Deaths (%)

Hazard Ratio (95 % confidence interval)

A1*A/*A

71

29(41)

1.0 (Reference)

A1*A/*B

92

34(37)

0.8(0.4-1.4)

A1*B/*B

33

7(21)

0.2 (0.1-0.6)

PlUe'^/lte™

110

35(32)

1.0 (Reference)

P1He10s/Val10s

107

33(31)

0.8(0.5-1.3)

105

10s

23

3(13)

0.3(0.1-1.0)

Mi-positive

133

46(35)

1.0 (Reference)

Mi-null

118

28(24)

0.6 (0.36-0.97)

T1-positive

177

58(33)

1.0 (Reference)

P1Val

/Val

T1-null

74

16(22)

0.5(0.29-0.91)

M1- and T1- positive

92

36(39)

1.0 (Reference)

M1- or T1- positive

126

32(25)

0.5 (0.31-0.86)

M1- and T1- negative

33

6(18)

0.3(0.12-0.77)

Data have been obtained for breast cancer patients treated for first incidence of breast cancer, for whom archival normal tissue and details of survival were available. Treatment always included cyclophosphamide as a first round of therapy followed by radiation or other treatments in some cases. The population is variable for race, stage at diagnosis, node status, estrogen receptor status and progesterone receptor status. The hazard ratios (for death) have been adjusted for these variables. Deaths are 'counted' no matter what cause (Data are taken from [31-34].) Genotypes that would be predictive of decreased hepatic detoxification of phosphoramide mustard (GSTA1'B/"B) or possibly predictive of decreased detoxification of 4-hydroxycyctophosphamide (GSTPf WWa/"*), and hence predictive of a more effective dose of chemotherapeutic agent, show increased survival (lower hazard ratio) compared to related 'higher detoxification" genotypes.

R F; CokJ atid F. F. Kadlitbar / Detoxi3('iccitnmc/fElectrophiiic Compounds

133

A putative model for the interaction of these genotypes with CP.metabolism is given in Fig. 8. We speculate that interindividual variation of the efficiency of CP chemotherapy involves activation in the liver, hepatic (GSTA 1genotype-dependent) detoxificationof PM, and GSTP1-dependent detoxification of 4-OH-CP in tumor and possibly in blood. GSTM1 and GSTTl may act via detoxification of products of oxidative stress associated with GSH depletion during combination chemotherapy. These studies need to be repeated in other populations and the validity of the model tested by studies of the pharmacogenetics of CP metabolism. However, there appears to be potential for the better design of chemotherapeutic regimens on the basis of individual GST (and other) genotypes. 5. GSTs and Homeostasis

The biological basis for the maintenance of organ-specific patterns of GST expression is not known. However, the conservation of phenotype suggests that, even though few endogenous

Fig. 8. Putative model for the interaction of GST genotypes and detoxtflcatlon of cyclophosphamlde (CP) metabolites as determlnants of chemotherapeutic efficiency of CP.

substrates are known [ 13, GSTs are of importance in cellular homeostasis by mechanisms other than detoxification of xenobiotics. The possibility exists that susceptibility to disease is associated with genetic variation of the GSTs because this affects cellular homeostasis per se rather than a direct response to toxic xenobiotics. This aspect of the GSTs has been less studied. GSTAl and GSTA2 catalyze the detoxification of fatty-acid hydroperoxides (including arachidonic acid hydroperoxide) and are, therefore, involved in protection fiom oxidative stress [l, 31. Other catalytic activities relevant to oxidative stress are given in Table 1. GSTPl has been shown to be involved in an oxidative stress-sensitive pathway, apparently operating independently of its catalytic properties, that determines the level of Jun N-terminal kinase activity, hence Jun activity and the apoptotic pathway [35, 361. It should also be noted that GSTPl regulation is Jun and Fos sensitive [37], establishing a potential feedback mechanism for GSTPl regulation and function. GSTA4 appears to be involved in the same pathway via its

134

B.F. Coles and F.F. Kadlubar / Detoxification of Electrophilic Compounds

efficient detoxification of 4-hydroxy-2-nonenal, a product of lipid peroxidation and a signal for apoptosis [38]. 6. Concluding Remarks In this discussion we have shown that, even within a single family of enzymes, there is a high degree of complexity of interaction between genetics, function and cancer etiology. We feel that the inconsistency of associations between GST genotype and susceptibility to cancer is a product of this complexity and the variability inherent in study populations. Because of this, we regard the associations that have been established as reliable evidence that the GSTs do have an effect on susceptibility to cancer in humans. In the case of chemotherapeutic agents (and potentially other therapeutic drugs) the role of GSTs may well be better defined, and this area of study has the potential to be of more immediate clinical importance. Fundamentally, tissue-specific expression, interindividual variability of expression and induction of GSTs (and other enzymes) will have a genetic basis that can be studied and used as a predictive tool. Methods for rapid, high throughput and multiple genotyping now exist in the form of DNA chips and mass spectral techniques. Similarly powerful methods exist for determination of phenotype via RNA expression, and are being rapidly developed for protein expression. Methods for estimating carcinogen exposure from diet (e.g., [39, 40]) and biologically relevant exposure using whole blood components [41] are also now advanced. The way forward is to build on the present generation of epidemiological studies by improving design to incorporate interactions between common risk factors, each of low penetrance when considered alone, but of greater attributable risk when considered in combination with exposure and high prevalence. In this way it should be possible to identify groups of risk factors that have real relevance to an individual, and which can form a sound basis for decisions concerning lifestyle, risk and management of disease.

References 1. J.D. Hayes, D.J. Pulford, The glutathione S-transferase supergene family: regulation of GST and the contribution of the isoenzymes to cancer chemoprotection and drug resistance, Grit

Rev. Biochem. Motec

Biol. 30(1995)445-600. 2. B. Coles, B. Ketterer, The role of glutathione transferases in chemical carcinogenesis, Grit. Rev. Biochem. Motec. Biol. 25 (1990) 47-70. 3. J. D. Hayes, L. I. McLellan, Glutathione and glutathione-dependent enzymes represent a co-ordinately regulated defence against oxidative stress, Free Rad. Res. 31 (1999) 273-300. 4. J.D. Hayes, R.C. Strange, Glutathione S-transferase polymorphisms and their biological consequences, Pharmacology 61 (2000) 154-166. 5. H.A.A.M. Dirven, B. van Ommen, P.J. van Bladeren, Glutathione conjugation of alkylating cytostatic drugs with a nitrogen mustard group and the role of glutathione S-transferases, Chem. Res.Toxicol. 9 (1996) 351-360. 6. S. Landi, Mammalian class theta GST and differential susceptibility to carcinogens: a review, Mutat. Res. 463 (2000) 247-283. 7. T.R. Rebbeck, Molecular Epidemiology of the human glutathione S-transferase genotypes GSTM1 and GSTT1 in cancer susceptibility, Cancer Epidemiol. Biomarkers Prevention, 6 (1997) 733-743.

B.F. Coles and F.F. Kadlubar / Detoxification of Electrophilic Compounds

135

8. A. Inskip, J. Elexperu-Camimaga, N. Buxton, P.S. Dias, J. Macintosh, D. Campbell, P.W. Jones, L. Yengi, J.A. Talbot, R.C. Strange, A.A. Fryer, Identification of polymorphism at the glutathione S-transferase GSTM3 locus: evidence for linkage with GSTM1*A, Biochem. J. 312 (1995) 713-716. 9. B.F. Coles, K.E. Anderson, D.R. Doerge, M.I. Churchwell, N.P. Lang, F.F. Kadlubar, Quantitative analysis of interindividual variation of glutathione S-transferase expression in human pancreas and the ambiguity of correlating genotype with phenotype, Cancer Res. 60 (2000) 573-579. 10. B.F. Coles, F. Morel, C. Rauch, W.W. Huber, M. Yang, C.H. Teitel, B. Green, N.P. Lang, F.F. Kadlubar, Effect of polymorphism in the human glutathione S-transferase A1 promoter on hepatic GSTA1 and GSTA2 expression, Pharmacogenetics 11 (2001) 663-669. 11. B. Coles, M.Yang, N.P. Lang, F.F. Kadlubar, Expression of hGSTPl alleles in human lung and catalytic activity of the native protein variants towards 1-chloro-2,4-dinitrobenzene, 4-vinylpyridine and (+)-anti benzo[a]pyrene7,8-diol-9,10-oxide, Cancer Lett. 156 (2000) 167-175. 12. N.Tetlow, D. Liu, P. Board, Polymorphism of human alpha class glutathione transferases, Pharmacogenetics 11 (2001)609-617. 13. J.D. Rowe, E. Nieves, I. Listowsky, Subunit diversity and tissue distribution of human glutathione Stransferases: interpretations based on electrospray ionization-MS and peptide sequence-specific antisera, Biochem. J. 325 (1997) 481-486. 14. A. van Ommen, J.J.P. Bogaards, W.H.M. Peters, B. Blaauboer, P.J. van Bladeren, Quantification of human hepatic glutathione S-transferases, Biochem. J. 269 (1990) 609-613. 15. W.H.M. Peters, L. Kock, F.M. Nagengasl, P.G. Kremers, Biotransformation enzymes in human intestine: critical low levels in the colon? Gut 32 (1991) 408-412. 16. C.L. Berendsen, W.H. Peters, P.G. Scheffer, A.A. Boumann, E. Boven, D.W. Newling, Glutathione Stransferase activity and subunit composition in transitional cell cancer and mucosa of the human bladder, Urology, 49 (1997) 644-651. 17. S. Anttila, L. Luostarinen, A. Hirvonen, E. Elovaara, A. Karjalainen, T. Nurminen, J.D. Hayes, H. Vainio, B. Ketterer, Pulmonary expression of glutathione S-transferase M3 in lung cancer patients; association with GSTM1 polymorphism, smoking, and asbestos exposure, Cancer Res. 55 (1995) 3305-3309. 18. S.S. Hecht, Tobacco smoke carcinogens and lung cancer, J. Natl, Cancer Inst. 91 (1999) 1194-1210. 19. F. Morel, O. Fardel, D.J. Meyer, S. Langouet, K.S. Gilmore, B. Meunier, C-P.D. Tu, T.W. Kensler, B. Ketterer, A. Guillouzo, Preferential increase of glutathione S-transferase class a transcripts in cultured human hepatocytes by phenobarbital, 3-methylcholanthrene and dithiolethiones, Cancer Res. 53, (1993) 231-234. 20. J.J.P. Bogaards, H. Verhagen, M.I. Willems, G. van Poppel, P.J. van Bladeren, Consumption of brussels sprouts results in elevated a-class glutathione S-transferase levels in human blood plasma, Carcinogenesis 15 (1994)1073-1075. 21. J.W. Lampe, C. Chen, S. Li, J. Prunty, M.T. Grate, D.E. Meehan, K.V. Barale, D.A. Dightman, Z. Feng, J.D. Potter, Modulation of human glutathione S-transferases by botanically defined vegetable diets, Cancer Epidemiol. Biomarkers Prevention 9 (2000) 787-793. 22. Y. Zhang, P. Talalay, Anticarcinogenic activities of organic isothiocyanates: chemistry and mechanisms, Cancer Res. 54 (1994) 1976s-1981s. 23. R.H. Kolm, H Danielson, Y Zhang, P. Talalay, B. Mannervik, Isothiocyanates as substrates for human glutathione S-transferases; structure-activity studies, Biochem. J. 311 (1995) 453-459. 24. K.R. Kaderiik, R.F. Minchin, G.J. Mulder, K.F. llett, M. Daugaard-Jenson, C.H. Teitel, F.F. Kadlubar, Metabolic activation pathway for the formation of DNA adducts of the carcinogen 2-amino-1-methyl-6-phenylimidazo[4,5bjpyridine (PhIP) in rat extrahepatic tissues, Carcinogenesis, 15 (1994) 1703-1709. 25. N.P. Lang, M.A. Butler, J. Massengill, M. Lawson, R.C. Stotts, M. Hauer-Jensen, F.F. Kadlubar, Rapid metabolic phenotypes for acetyltransferase and cytochrome P4501A2 and putative exposure to food-borne heterocyclic amines increase the risk for colorectal cancer or polyps, Cancer Epidemiol. Biomarkers Prevention 3 (1994) 675-682.

136

B.F. Coles and F.F. Kadlubar / Detoxification of Electrophilic Compounds

26. B. Coles, S.A. Nowell, S.L MacLeod, C. Sweeney, N.P. Lang. F.F. Kadlubar, The rote of human glutathione Stransferases (hGSTs) in the detoxification of the food-derived carcinogen metabolite N-acetoxy-PhIP, and the effect of a polymorphism in hGSTAn on cotorectal cancer risk, Mutat. Res. 482 (2001) 3-10. 27. D.W. Layton, K.T. Bogen, M.G. Knize, FT. Hatch, V.M. Johnson, J.S. Fetton, Cancer risk of heterocydic amines in cooked foods: an analysis and implications for research, Carcinogenesis 16 (1995) 39-52. 28. L. Le Marchand, J. H. Hankin, LR. Wilkens, LM. Pierce, A. Franke, LN. Kolonel. A. Seifried, L.J. Ouster, W. Chang, A. Lum-Jones, T. Donlon, Combined effects of well-done red meat, smoking, and rapid Nacetyttransferase 2 and CYP1A2 phenotypes in increasing cotorectal cancer risk, Cancer Epidemtol. Bksmarkers Prevention 10 (2001) 1259-1266. 29. A. Sweeney, B.F. Coles, S. Nowell, N.P. Lang, F.F. Kadlubar, Novel markers of susceptibility to carcinogens in diet: associations with cotorectal cancer, Toxicology, (2002) in press. 30. H.J. Lin, N.M. Probst-Hensch, A.D. Louie. I.H. Kau, J.S. Witte, S.A. Ingles, H.D. Frank), E.R. Lee, R.W. Haite, Glutathione transferase null genotype, broccoli, and lower prevalence of cotorectal adenomas. Cancer Epidemiol. Btomarkers Prevention 7 (1998) 647-652. 31. C. Sweeney, C.B. Ambrosone, L. Joseph, A. Stone, LF. Hutchins, F.F. Kadlubar, B.F. Cotes. Association between a glutathione S-transferase A1 promoter polymorphism and survival after breast cancer treatment, Cancer Res. (2002), in press. 32. C. Sweeney, G.Y. McClure, M.Y. Fares, A. Stone, B.F. Cotes, P.A. Thompson, S. Korourian, LF. Hutchins, F.F. Kadlubar, C.B. Ambrosone, Association between survival after treatment for breast cancer and glutathione S-transferase P1 ItelOSVal polymorphism, Cancer Res. 60 (2000) 5621-5624. 33. C. Sweeney, L. Joseph, B.F. Cotes, G.Y. McClure, P.A. Thompson, M.Y. Fares, LF. Hutchins, F.F. Kadlubar. C.A. Ambrosone, Glutathione S-transferase (GST) P1 expression in primary breast tumors in relation to GSTP1lle105Val genotype and survival, Proc. Amer. Assoc. Cancer Res., 42 (2001) 739. 34. C.B. Ambrosone, C. Sweeney, B, Cotes, P.A. Thompson, G.Y. McClure, S. Korourian, M.Y. Fares, A. Stone, F.F. Kadlubar, LF. Hutchins, Polymorphisms in glutathione S-transferases (GSTM1 and GSTT1) and survival after treatment for breast cancer. Cancer Res. 61 (2001) 7130-7135. 35. V. Adler, Z. Yin, S.Y. Fuchs, M. Benezra, L. Rosario, K.D. Tew, M.R. Pincus, M. Sardana. C.J. Henderson, C.R. Wolf, R.J. Davis, Z. Ronai, Regulation of JNK signaling by GSTp, EMBO J. 18 (1999) 1321-1334. 36. A. Villafania, K. Anwar, S. Amar, L. Chie, D. Way, D.L Chung, V. Adler, Z. Ronai, P.W. Brandt-Rauf. Z. Yamaizumii, H-F. Kung, M. R. Pincus, Glutathione S-transferase as a selective inhibitor of oncogenic ras-p2linduced mrtogenic signaling through blockade of activation of jun by jun-N-terminal kinase. Ann. Clin. Lab. Sci., 30 (2000) 57-64. 37. C.J. Henderson, A.W. McLaren. G.J. Moffat, E.J. Bacon, C.R. Wolf, Pi-class glutathione S-transferase: regulation and function, Chemico.-Btol. Interact. 111-112 (1998) 69-82. 38. J-Z. Cheng, S.S. Singhal, A. Sharma, M. Saini, Y.Yang, S. Awasthi, P. Zimniak, Y.C. Awasthi, Transfectton of mGSTA4 in HL-60 cells protects against 4-hydroxynonenal-induced apoptosis by inhibiting JNK-mediated signaling, Arch. Biochem. Biophys. 392 (2001) 197-207. 39. G.A. Keating, R. Sinha, D. Layton, C.P. Salmon, M.G. Knize, K.T. Bogan, C.F. Lynch. M. Alavanj, Comparisons of heterocyclic amine levels in home-cooked meats with exposure indicators, Cancer Causes Control 11 (2000) 731-739. 40. N. Kazerouni, R. Sinha, C.H. Hsu, A. Greenberg, N. Rothman, Analysis of 200 food items for benzo[a]pyrene and estimation of its intake in an epidemtologic study, Food Chem. Toxicol. 39 (2001) 423-436. 41. S. Pavanelto, E. Ctonfero, Biological indicators of genotoxic risk and metabolic polymorphisms, Mutat. Res. 463 (2000) 285-308. 42. A-S. Johansson, B. Mannen/ik, Human glutathione transferase A3-3, a highly efficient catalyst of doubte-bond isomerisation of the biosynthetic pathway of steroid hormones, J. Btol. Chem. 276 (2001) 33061 -33065. 43. S.K. Srivastava, S.S. Singhal, X. Hu, Y.C. Awasthi, P. Zimniak, S.V. Singh, Differential catalytic efficiency of altelic variants of human glutathione S-transferase Pi in catalyzing the glutathione conjugation of thtotepa. Arch. Biochem. Biophys. 366 (1999) 89-94.

B.F. Coles and F. F. Kadlubar / Detoxification of Elect rophilic Compounds

137

44. G.A. Toriiner, C. Akyerli, A. Ucar, T. Aki, N. Atsu, H. Ozen, M. Tez, M. Cetinkaya, T. Ozcelik, Polymorphisms of glutathione S-transferase genes (GSTM1, GSTP1 and GSTT1) and bladder susceptibility in the Turkish population, Arch. Toxicol. 75 (2001) 459-464. 45. A. Risch, H. Wikman, S. Thiel, P. Schmezer, L. Edler, P. Drings, H. Dienemann, K. Kayser, V. Schulz, B. Spiegelhajder, H. Bartsch, Glutathione S-transferase M1, M3, T1 and P1 polymorphisms and susceptibility to non-small-cell lung cancer subtypes and hamartomas, Pharmacogenetics 11 (2001) 757-764. 46. K. Mitrunen, N. Jourenkova, V. Kataja, M. Eskelinen, V.M. Kosma, S. Benhamou, H. Vainio, M. Uusitupa, A. Hirvonen, Glutathione S-transferase M1, M3, P1 and T1 genetic polymorphisms and susceptibility to breast cancer, Cancer Epidemiol. Biomarkers Prevention 10 (2001) 229-236. 47. C.M. Maugard, J. Charrier, A. Pitard, L. Campion, O. Akande, L. Pleasants, F. Ali-Osman, Genetic polymorphism at the glutathione S-transferase (GST) P1 locus is a breast cancer risk modifier, Int. J. Cancer 91 (2001)334-339.

138

Thiol Metabolism and Redox Regulation of Cellular Functions A. Pompella el ai (Eds.) IOS Press, 2002

Transcriptional Regulation of Glutathione S-Transferase Pl-1 in Human Leukemia A.DUVOIX, M. SCHMITZ, M. SCHNEKENBURGER, M. DICATO, F. MORCEAU, M.-M. GALTEAU and M. DIEDERICH Laboratoire RCMS, Centre Universitaire du Luxembourg Batiment des Sciences, 162A Avenue de la Faiencerie, L-1511 Luxembourg, Luxembourg Tel: (+ 352) 46 66 44 434 - Fax: (+ 352) 46 66 44 438 - Email: [email protected] 1. Introduction The glutathione S-transferases (GST) (EC 2.5.1.18) are a multigene superfamily of enzymes that catalyze the conjugation with glutathione of a number of electrophilic compounds including xenobiotic drugs, toxins and carcinogens, as well as some endogenous cellular electrophiles [1, 2]. GST are believed to play a key role in the protection of cells from the toxicities of xenobiotic compounds, as well as from lipid hydroperoxides generated by oxidative stress [3]. Mammalian GSTs have been classified into nine distinct gene families: seven cytosolic groups (alpha, mu, pi, theta, omega, kappa and zeta), one microsomal form and one form present in erythrocytes. GSTa is mainly expressed in the liver and the kidney while GSTP1-1 is expressed as a major form in organs such as lung, breast or bladder [4, 5]. In many human tumors, like prostate carcinoma [6], squamous-cell carcinoma [7], acute lymphoblastic leukemia [8] and chronicle lymphoid leukemia [9], GSTP1-1 is overexpressed, even though in the corresponding normal tissues the protein is either absent or expressed at very low levels. GSTP1-1 can thus be used as a valuable prognostic tool in sarcoma [10] or gastric carcinoma [11]. GSTP1-1 appears to be involved in the development of anticancer drug resistance, and elevated levels of GSTP1 mRNA are found in cell lines resistant to a range of anticancer drugs. Indeed, MCF7, an oestrogen-receptor positive breast cancer cell line, was found to overexpress GSTP1-1 and to develop a resistance to ethacrynic acid [12], doxorubicin and benzopyren [13]. Other cell lines that also overexpress GSTP1-1 are resistant to doxorubicin or taxol [14]. COS cells becomes resistant to doxorubicine [15] and CHO cells resist to cisplatine and carboplatine [16] after transfection of the GSTP1-1 gene. 2. GSTP1-1 in Human Leukemia Wang et al. (2000) [17] studied the expression and the activity of GST isoenzymes in 14 haematopoietic lineages. They found that GSTP1 expression was higher than other GSTs in 13 out of 14 cell lines and they established the best correlation between GSTP1 expression and l-chloro-2,4-dinitrobenzene (CDNB) conjugation activity in U937, K562 and Jurkat cells. Increased levels of GSTP1-1 are associated with tumor development and carcinogenesis [6]. Sauerbrey et al. (1994) [8] have shown elevated GSTP1-1 levels in the blasts from childhood acute lymphoblastic leukemia; moreover their data show an association between increased GSTP1-1 levels and a higher relapse rate, as well as a lower probability of the first continuous complete remission. The expression of GSTP1-1 in acute

A. Duvoix et al. / Transcriptional Regulation of Glutathione S-Transferase

139

non-lymphoblastic leukemia cells is significantly correlated with response to induction therapy, duration of first remission, and overall survival of the patients [18]. However, a predictive correlation with tumor development has not been conclusively established, as Marie et al. (1995) [19] have shown a decrease in GSTPl gene expression in mononuclear cells of patients with chronic lymphocytic leukemia. An involvement of GSTPl in the development of resistance to some antineoplasic drugs and genotoxic carcinogens has also been implicated in cell lines selected for resistance to a range of anticancer drugs. However the molecular mechanisms responsible for changes in GSTPl expression are poorly understood.

3. Comparison of the Human and Rat GSTP Promoters Detailed studies of the rat P class GST (GSTP) promoter have identified several regulatory elements necessary for basal and inducible expression. Analyses of the 5'-flanking region of the human GSTPl gene have suggested that transcriptional regulation of GSTPl differs significantly from the rat GSTP homologue. The rat also expresses a GSTP that is not present in liver, however, unlike GSTPl-1 in human, the enzyme expression is increased during hepatocarcinogenesis [20]. This difference is due to the structure of the rat GSTP gene promoter (Figure 1). This promoter contains two enhancers (GPEI and GPEII), a silencer (GPS1), a GC box and a TRE [21]. GPEI, a strong enhancer, is composed of two non-consensus TRE and mediates GSTP expression. Each TRE alone is inactive but synergistically cis-activate GSTP. Their activity is orientation dependant [22], and the upstream 19 nucleotides are essential for the maximal activity of the enhancer [23]. Its basal activity is not due to the binding of AP-1 transcription factors, although it can bind to the downstream TRE and increase GSTP expression [24, 25]. Other factors, not yet known, bind to GPEI to activate it. Some of those factors are not found in normal liver but are present in hepatocarcinoma, and that could explain the increase expression of GSTP in those cells [26]. GPEn is a weak enhancer composed of two SV40 and one polyoma enhancer. The TRE, close to the site of initiation of the transcription, could also be part of the regulation system as its deletion can strongly decrease GSTP expression [27]. GPS1 is a silencer and is position and orientation independent [28]. This region binds several factors (Silencer Factor (SF) -A, -B, -C). SF-B is part of the family CCAAT/enhancer-binding protein (C/EBP). In normal liver cells, C/EBPa binds to the silencer. In carcinoma cells, the ratio C/EBPa / C/EBPp decrease and C/EBPP binds to GSP1 increasing the GSTP expression. That could be another explanation for the increased GSTP activity found in hepatocarcinoma [29]. SF-A is part of the Nuclear Factor (NF) -1 family and contributes to the silencing of GSTP [30]. For the human gene, Morrow et al (1990) [2] reported the localization of a region spanning from -80 to -8, which is absolutely required for reporter gene activity in transient transfection experiments. Moreover a region from -73 to +8, as shown by Xia et al.(1996) [31], is absolutely required for retinoic acid dependent repression. Other findings suggest that the activator protein 1 site (API, located from -69 to -63) is essential for promoter activity [32]. Analysis of the GSTPl proximal promoter also revealed the presence of two putative Spl binding sites located downstream of the AP-1 response element (-57 to -49 and -Al to -39). According to Jhaveri et al (1998) [33], the proximal G/C box is essential for promoter activity whereas the distal box is not required, as shown by site directed mutagenesis. Moreover, a putative nuclear factor kB (NFkB) binding site has been localized at positions -96 to -86 by Xia et al(1996) [31].

A. Duvoix et al. / Transcriptional Regulation of Glutathione S-Transferase

140

-2000

-3000

-1000

NF-kB-like -93

iPEI-

Spl -56

TRE NFE2 -63

GPSI

TRE

Spl -48

GCbox

TCGATAGTCAGTCACTATGATTCAGCAATCG ^ k. A TRE like TRE like

ARE/EpRE -700 C/EBP like

TRE like

TRE like

Fig. 1. Organisation of the human GSTP1-1, the rat GSTP and the rat GST Ya promoters. NF-kB-like: Nuclear Factor kB-like, TRE: TPA response element, GPE: GSTP enhancer, GPS: GSTP silencer, AhRE: Ah response element, ARE/EpRE: antioxidant response element / electrophile response element.

4. Regulation of GSTP Expression by Activating Protein 1 Protein components of activating protein 1 (AP-1) are encoded by a set of genes called "immediate-early genes" whose transcription is rapidly induced, independently ofde novo

A. Diivoix et al. / Transcriptional Regulation of Glutathione S-Transferase

141

protein synthesis. The AP-1 transcription factor is a complex composed of proteins of the fos and jun proto-oncogene families, which need to dimerize to promote binding of the complex to the AP-1 recognition site. Moffat et a/.(1994) [32] have already shown, in a human breast carcinoma cell line, the binding of Jun and Fos proteins to the AP-1 site located in the GSTP1 promoter. The basal transcription of the human NAD(P)H:quinone oxidoreductasel gene, another drug metabolizing enzyme, is mediated by antioxidant response elements (ARE) containing perfect AP-1 sites or TRE-like elements. Xie et al. (1995) [34] demonstrated that ARE (composed of two TRE or TRE-like elements)-containing detoxifying enzymes genes are responsive to xenobiotics and antioxydants. JunD and c-Fos proteins, which are present in the nuclear extracts derived from hepatic cells, have been shown to bind to the AP-1 site located in the promoter of the human NAD(P)H:quinone oxidoreductasel (NQOl)gene[35]. Exposure of murine hepatoma cells to chemical inducers of GST Ya gene expression such as phorbol esters was found to induce an increase in AP-1 binding activity. This increase was shown to involve the induction of fos and jun gene expression with accumulation of increased levels of the respective mRNA and a de novo synthesis of the AP-1 protein components [36].

5. Methylation of the GSTP Promoter In addition to transacting factors, epigenetic events like CpG island methylation near the GSTP1 gene might be of importance for the regulation of GSTP 1 expression. Such CpG island methylations close to the GSTP1 locus accompany development of human breast [37] and prostatic carcinomas [38]. A better understanding of the molecular basis of the expression of glutathione metabolizing enzymes is crucial for the understanding of the protecting role against alkylating agents of glutathione in human chronic leukemia. Involvement of DNA methylation in the lineage specific gene regulation in haematopoietic cells has already been demonstrated in calcitonin gene hypermethylation in chronic myeloid leukemia [39] and hypomethylation of the major bcr gene in Ph positive acute lymphoid leukemias [40], In a recent study [41] we have examined the relationship between methylation and the promoter activity of a 136-bp minimal GSTP I promoter in the human leukemia cell line K562. We found that in vitro methylation of this promoter with Sssl has a transcriptional inhibitory effect and that the methylation state of this GSTP1 promoter fragment in expressing and non-expressing leukemia cell lines correlates with the degree of GSTP1 RNA expression and transcriptional activity (Figure 2). 6. Molecular Mechanisms of GSTP1-1 Expression in Human Leukemia Many genes are responsive to the transcriptional activation of TPA via the recognition of a consensus AP-1 binding site by trans-acting factors such as Jun and Fos [42, 43]. We showed recently that TPA activation of human leukemia K562 cells produces an increase in the level of GSTP1 mRNA [44]. Our results were in conflict with previous reports that showed the absence of effect of TPA on GSTP1 mRNA in HeLa cells [2] or in HepG2 cells [45]. For these authors, their results were paradoxical as they showed that the integrity of the AP-1 binding site is essential for the basal activity of the GSTP1 promoter [42]. The reason for the discrepancy between our results and previous data is unknown, but may be related to the different types of cells used. In agreement with that hypothesis, TPA was shown to be effective on the expression of murine GST Ya [36] or rat GST P [21], whose

142

A. Duvoix et al. / Transcriptional Regulation of Glutathione S-Transferase

promoters contain two AP-1 binding sites. Moreover, TPA was shown to increase the binding of AP-1 factors to the promoters of erythroid-specific genes in the K562 cells [46].

A) Normal prostate cells

Expressing GSTP1-1 A1AAA i9_24 -507 -410

t

t

Hypermethylated region

Hypomethylated region

B) Prostate "Cancer cells

Not expressing GSTP1-1

Expressing GSTP1-1

-97

t Partially demethylated region Not expressing GSTP1-1 GSTP1-1

Fig. 2. Methylation state of GSTP1 gene promoter in different cell human tissues or cell lines. A) Normal prostate cells (expressing GSTP1-1): hypermethylated region upstream of -500 bp, and hypomethylated region between +1 and -400 bp. B) Prostate cancer cells (not expressing GSTP1-1): fully hypermethylated promoter. C) K562 cell line (expressing GSTP1-1): partially demethylated promoter in the region +1 -97. D) Raji lymphoma cell line (not expressing GSTP1-1): region +1 -97 is fully methylated.

To understand the molecular mechanism of the transcriptional induction of GSTP1-1 in response to TPA, we cloned the 5'-flanking region of the GSTPJ into a luciferase reporter system. Following transfection in K562 cells, we observed that TPA up-regulated the promoter region of the GSTPJ gene, measured as increased luciferase activity. This suggests that TPA acts at the level of transcription to induce GSTPJ mRNA in human leukemia cells. We next elucidated the transcriptional regulatory mechanism by which TPA exerted its effect on the induction of GSTPJ. Some investigators have shown the critical

A. Duvoix et al. / Transcriptional Regulation of Glutathione S-Transferase

143

role of the AP-1 recognition sequence for transcriptional activity of the GSTPJ promoter [32, 33]. However they have suggested that the distal Spl site and a putative NF-kB like site [31] may be important in regulating GSTP1 gene expression. We therefore used TPA to assess the role of the AP-1 transcription factor in the transcriptional up-regulation of GSTPJ [44]. Exposure of leukemia cells to TPA produced a significant increase in the DNA binding activities of nuclear proteins. The competition studies confirmed the activation of AP-1 by TPA in our cells. To identify which components of AP-1 are responsible for the up-regulation of GSTPJ, we used antibodies directed against c-Fos and c-Jun families of proteins. Antibodies that cross-reacted with c-Jun and Fral produced a supershift. In support of this, several investigators have suggested possible involvement of AP-l/Jun family members in the regulation of GSTPJ. Jhaveri and Morrow (1998) [33] failed to demonstrate the presence of transcription factors belonging to the Jun family in the protein complexes bound on a TRE probe with nuclear extracts derived from MCF7 or HS578T cells. Dixon et al. (1989) [45] and Morrow et al. (1989) [47] had already shown that the AP-1 site in the GSTPJ promoter was unresponsive to the Jun and Fos trans-acting factors in HeLa and FfepG2 cells. But they did not use supershift experiments to obtain these results; cells were co-transfected with v-jun, c-jttn or c-fos expression vectors, alone or in combination. In contrast, Moffat et al. (1994) [32] demonstrated, by using specific antisera, that Jun and Fos proteins are integral components of the VCREMS nuclear complex bound on the AP-1 site; however these specific effects of antisera were not observed with MCF7 cells. Their results mean that the same promoter region of GSTP1 is able to bind members of the Jun and Fos protein families in some cell lines but not in all the cells.

7. Conclusion Our leukemia model might be of particular interest for the better understanding of the expression mechanisms of the GSTP1-1 gene, a human gene specifically related to drug resistance. Our findings contribute to a better understanding of the molecular mechanisms involved in GSTP1 expression which are of major importance for the development of chemoresistance reducing the impact of therapeutic agents in human haematopoietic diseases.

Acknowledgments This work was supported by the Fondation de Recherche "Cancer et Sang". AD, MS, and MSch were supported by fellowships from the Government of Luxembourg. The authors also thank the Action Lions "Vaincre le Cancer", "Les Amis de la Fondation Jose Carreras - Luxembourg" and the "CNFPC / Esch-sur-Alzette" for supporting this project.

References [1] D.L Schipper, Glutathione S-Transferases and cancer, Int. J. Oncol., 10 (1997) 1261-1264. [2] C.Morrow , Regulation of human glutathione S-transferase pi gene transcription : influence of 5'flanking sequences and transactivating factors which recognize AP-1 binding sites, Gene, 88 (2) (1990) 215-225. [3] Y.Yang, Role of glutathione S-transferase in protection against lipid peroxidation. Overexpression of hGSTA2-2 in K562 cells protects against hydrogen peroxide-induced apoptosis and inhibits JNK and caspase 3 activation, J. Biol. Chem., 276(22) (2001) 1922019230.

144

A. Duvoix et al. / Transcriptional Regulation of Glutathione S-Transferase

[4] P.G.Board, Biochemical genetics of glutathione-S-transferase in man, Am J Hum Genet, 33(1) (1981) 36-43. [5] J. Commandeur, Enzymes and transport systems involved in the formation and disposition of glutathione S-conjugates, Pharmacol. Rev., 47 (2) (1995) 271-330. [6] M.S. Cookson, Glutathione S-transferase pi dass expression by immunohistochemistry in benign and malignant prostate tissue, J. Urol. 157(1997) 673-676. [7] T. Inoue, Glutathione S-transferase pi is a powerful indicator in chemotherapy of human lung squamous-cell carcinoma, Respiration. 62 (1995) 223-227. [8] A.Sauertrey, P-glycoprotein and glutathione S-transferase IT in childhood acute lymphoblastic leukemia, Br. J. Cancer, 70 (1994) 1144-1149. [9] J.C. Schisselbauer, Characterisation of glutathione S-transferase expression in lymphocytes from chronic lymphoblastic leukemia cancer, Cancer Res., 50 (1990) 3562-3568. [10] G. Toffoli, Expression of MDR1 and GST-pi in human soft tissue sarcomas: relation to drug resistance and biological aggressiveness, Ann Oncol, 3(1) (1992) 63-69. [11] N. Monden, Prognostic significance of the expressions of metallothionein, glutathione-Stransferase-pi, and P-glycoprotein in curatively resected gastric cancer, Oncology, 54(5) (1997) 391-399. [12] C.S. Morrow, Combined expression of multidrug resistance protein (MRP) and glutathione Stransferase P1-1 (GSTP1-1) in MCF7 cells and high level resistance to the cytotoxicities of ethacrynic acid but not oxazaphosphorines or cisplatin, Biochem Pharmacol. 56(8) (1998) 10131021. [13] J.A. Moscow, Elevation of TT dass glutathione transferase activity in human breast cancer ceNs by transfedion of the GSTir gene and its effect on sensitivity to toxins, Molecular Pharmacology, 36(1989)22-28. [14] U. Masanek, Messenger RIMA expression of resistance proteins and related factors in human ovarian carcinoma cell lines resistant to doxorubicin, taxol and cisplatin, Anticancer Drugs, 8(2) (1997) 189-198. [15] R.B. Puchalski and W.E. FaM, Expression of recombinant glutathione S-transferase TT, Ya, or Yb1 confers resistance to alkytatjng agents, Medical Sciences. 87 (1990) 2443-2447. [16] M. Miyazaki, Drug resistance to ds-diamminedichloroplatinum (II) in Chinese hamster ovary cell lines transfeded with glutathione S-transferase pi gene, Biochem Biophys Res Commun. 166(3) (1990) 1358-1364. [17] L. Wang, Glutathione S-transferase enzyme expression in hematopoietic cell lines implies a differential protective role for T1 and A1 isoenzymes in erythroid and for M1 in lymphoid lineages, Haematologica, 85 (2000) 573-579. [18] U.Tidefelt, Expression of glutathione transferase TT as a predictor for treatment result at different stages of acute nonlymphoblastic leukemia, Cancer Res., 52 (1992) 3281-3285. [19] J.P. Marie, Glutathione S-transferase and mdrl mRNA expression in normal lymphocytes and chronic lymphocytic leukemia, Leukemia, 9 (1995) 1742-1747. [20] A.Krtahara, Changes in molecular forms of rat hepatic glutathione S-transferase during chemical hepatocarcinogenesis, Cancer Res., 44(6) (1984) 2698-2703. [21] M. Sakai, multiple regulatory elements and phorbol 12-O-tetradecanoate 13-acetate responsiveness of the rat placenta! glutathione transferase gene, Proc.NatJ.Acad.Sci., 85(24) (1988) 9456-9460. [22] A. Okuda, Functional cooperativity between two TPA responsive elements in undifferentiated F9 embryonic stem cells, 9(4) (1990) 1131-1135. [23] A. Okuda, Structural and functional analysis of an enhancer GPEI having a phorbol 12-Otetradecanoate 13-acetate responsive element-like sequence found in the rat glutathione transferase P gene, J. Biol. Chem.. 264(28) (1989) 16919-16926. [24] M.B. Diccianni, The dyad palindromic glutathione transferase P enhancer binds multiple factors including AP-1, Nudeic Acid Res., 20(19) (1992) 5153-5158. [25] M. Sakai, Suppression of glutathione transferase P expression by glucocottjcoid, Biochem. Biophys. Res. Commun., 187(2) (1992) 976-983. [26] D. Liu, Effect of trans-acting factors on rat glutathione S-transferase P1 gene transcription regulation in tumor cells, Chin. Med. J., 115(1) (2002) 103-106. [27] T. Suzuki, Tissue-specific activation of tumor marker glutathione transferase P transgenes in transgenic rats, J. Cancer Res. Clin. Oncol., 121(9-10) (1995) 606-611. [28] M. Imagawa, Silencer binding proteins function on multiple cis-element in the glutathione transferase P gene, Nudeic Acid Res., 19(1) (1991) 5-10. [29] S. Osada, CCAAT/enhancer-binding proteins alpha and beta interact with the silencer element in the promoter of glutathione S-transferase P gene during hepatocarcinogenesis, J. Biol. Chem., 270(52) (1995) 31288-31293.

A. Duvoix et al. / Transcriptional Regulation of Glutathione S-Transferase

145

[30] S. Osada, Nuclear factor A family proteins bind to the silencer element in the rat glutathione transferase P gene, J. Biochem., 121(2) (1997) 355-363. [31] C. Xia, The organisation of the human GSTP1-1 gene promoter and its response to retinoic acid and cellular redox status, Biochem. J, 313 (1996) 155-161. [32] G.J. Moffat, Involvement of Jun and Fos proteins in regulating transcriptional activation of the human pi class glutathione S-transferase gene in muttidrug-resistant MCF7 breast cancer cells. J. Biol. Chem., 269 (1994) 16397-16402. [33] M. Jhaveri and C.Morrow, Contribution of proximal promoter elements to the regulation of basal and differential glutathione S-transferase P1 gene expression in human breast cancer cells, Biochim. Biophys. A eta, 1396 (2) (1998) 179-190. [34] T.Xie, ARE- and TRE-mediated regulation of gene expression, J. Biol. Chem., 270 (12) (1995) 6894-6900. [35] Y.Li and AJaiswal, Identification of jun-B as third member in human antioxidant response element-nuclear proteins complex. Biochem. Biophys, Res. Comm., 188 (1992) 992-996. [36] R.Pinkus, Role of oxydants and antioxydants in the induction of AP-1, NF-kB, and glutathione S-transferase gene expression, J. Biol. Chem., 271 (23) (1996) 13422-13429. [37] M.S.Jhaveri and C.S.Morrow, Methylation-mediated regulation of the glutathione S-transferase P1 gene in human breast cancer cells, Gene, 210 (1998) 1-7. [38] W.H. Lee, Cytidine methylation of regulatory sequences near the it-class glutathione Stransferase gene accompanies human prostatic carcinogenesis, Proc. Natl. Acad. Sci., 91 (1994)11733-11737. [39] J.R.Melki, Concurrent ONA hypermethylation of multiple genes in acute myeloid leukemia, Cancer Res., 59 (1999) 3730-3740. [40] J.H.Ohyashiki, The methylation status of the major breakpoint cluster region in human leukemia cells, including Philadelphia chromosome-positive cells, is linked to the lineage of hematopoietic cells, Leukemia, 7(6) (1993) 801-807. [41] P. Borde-Chiche, Regulation of transcription of the glutathione S-transferase P1 gene by methylation of the minimal promoter in human leukemia cells, Biochem. Pharmacol., 61(5) (2001) 605-612. [42] L.Xia, Glutathione transferase pi: its minimal promoter and downstream cis-acting element, Biochem. Biophys. Res. Com., 176 (1991) 233-240. [43] G.Cirillo, Role of distinct mitogen-activated protein kinase pathways and cooperation between Ets-2, ATF-2, and Jun family members in human Urokinase-type plasminogen activator gene induction by interieukin-1 and tetradecanoyl phorbol acetate, Mol. Cell. Biol,, 19 (9) (1999) 62406252. [44] P. Borde-Chiche, Phorbol ester responsiveness of the glutathione S-transferase P1 gene promoter involves an inducible c-jun binding in K562 leukemia cells, Leuk. Res., 25(3) (2001) 241-247. [45] K.M. Dixon, Control of expression of the human glutathione S-transferase TT gene differs from its rat orthologue. Biochem. Biophys, Res. Comm., 163 (2) (1989) 815-822. [46] W.B.Solomon, Suppression of a cellular differentiation program by phorbol esters coincides with inhibition of binding of a cell-specific transcription factor (NF-E2) to an enhancer element required for expression of an erythroid-specific gene, J. Biol. Chem., 268 (7) (1993) 5089-5096. [47] C.S.Morrow, Structure of the human genomic glutathione S-tranferase pi gene. Gene, 75 (1989)3-11.

146

Thiol Metabolism and Redox Regulation of Cellular Functions A. Pompella et al. (Eds.} IOS Press, 2002

Mechanism of y-Glutamyltranspeptidase Folding and Activation in the Endoplasmic Reticulum Rebecca P. HUGHEY University of Pittsburgh School of Medicine, Department of Medicine, Renal-Electrolyte Division and Dept. of Molecular Genetics and Biochemistry-Pittsburgh, PA 15213, USA

1. Oxidative Stress in the Kidney The kidney is a target of oxidative stress from many sources including exposure to toxic compounds, dietary deficiency of antioxidants, diabetes mellitus, invasion by immune cells, or ischemia and re-perfusion (reviewed in [1, 2]). Reactive oxygen species (ROS) such as hydrogen peroxide (H2O2), hydroxyl radical (OH») and superoxide radical (O/) are the usual source of oxidative damage that results from a failure of normal cellular antioxidant mechanisms to cope with either their normal production or excess accumulation. The role of ROS in post-ischemic renal injury is based on finding subsequent lipid peroxidation in renal tissue and its prevention by prior infusion of an antioxidant enzyme like superoxide dismutase or the small metabolite glutathione in its reduced form. This protective role for glutathione against hypoxic injury has also been observed in model systems of proximal tubule suspensions or cultures such as LLC-PKicells from pig kidney.

2. Glutathione Protects against Oxidative Stress Glutathione is the major non-protein thiol of the body and is synthesized in the cytoplasm of cells in the reduced form (GSH) [3,4]. It is a cofactor for numerous enzymatic reactions involved in reduction of ROS, producing the oxidized form of glutathione (GSSG), and for detoxification of electrophilic substances, often by forming mixed disulfides (mercapturic acids). The importance of this antioxidant function is consistent with finding a 50:1 ratio of GSH/GSSG in the cytosol. However, this ratio is quite different in the endoplasmic reticulum (ER) where it is about 1:1, a unique redox state that is very similar to that which is used in vitro for optimal protein disulfide formation. This information led to the proposal that glutathione-mediated redox buffering in the ER was central to the formation of de novo protein disulfides by the ER chaperone, protein disulfide isomerase (PDI) [5,6]. However, more recent studies implicate oxidized Erolp, a protein first discovered in yeast, as the key component of the pathway for oxidizing protein thiols within the ER [7,8,9]. Erolp uses molecular oxygen as the primary donor of oxidizing equivalents but the exact mechanism is still unknown. In addition, these studies show that normal protein disulfide formation can occur in the absence of glutathione and is antagonized by glutathione excess. Taken together, this suggests that glutathione is competing with protein thiols for interaction with PDI. However, when cells were stressed with hyperoxidizing conditions by pretreatment with diamide, normal protein disulfide formation was only found in the presence of normal levels of glutathione.

R.P. Hughey / y-Glutamyltranspeptidase Folding and Activation

147

This suggests that one role of glutathione in the ER is to protect against hyperoxia by acting as a redox buffer.

3. Turnover of Glutathione Both GSH and GSSG are released from many tissues into the blood and more than 80% is extracted in a single pass through the kidney [3]. The renal extraction of GSH/GSSG results from its hydrolysis by the ectoenzyme y-glutamyltranspeptidase (yGT) that releases glutamate and cysteinylglycine (reduced or oxidized, respectively). The latter is then cleaved to cysteine/cystine and glycine by the action of aminopeptidase N or dipeptidase. All three of these enzymes are localized primarily at the apical surface of the renal proximal tubule while a smaller, but significant percent, is on the basolateral surface. Renal recovery of individual amino acids utilizes the sodiumdependent transport systems that are associated with the plasma membrane. Although transporters for intact glutathione are described on the plasma membrane, the efficient degradation of glutathione at both the apical and basolateral surface suggests that this electrogenic system is actually involved in secretion rather than reabsorption of glutathione. We have previously shown that the turnover of intracellular glutathione in polarized cultures of LLC-PKi cells grown on permeable filters results from similar rates of secretion at the apical and basolateral surface; this can be observed only when yGT-mediated degradation was inhibited with acivicin (AT-125) [10]. More recent studies by others [11] indicate that the secretion of glutathione from either J558L myeloma cells orXenopus laevis oocytes is blocked with inhibitors of vesicular traffic like brefeldin A or monensin. More notably, the redox state of the secreted glutathione is directly proportional to the disulfide content of any given transfected protein. Thus, expression of proteins enriched in cysteine disulfides, such as the immunoglobulin A, chain or influenza hemagglutinin (HA), causes both an increase in glutathione secretion and a 10-fold increase in the thiolrdisuffide ratio (GSH 4- cysteine : GSSG + cystine). Control proteins that lack disulfides do not have this effect. Thus, GSSG may serve as a reserve of oxidative equivalents for PDI within the ER when the Erolp-dependent pathway is saturated or stressed.

4. Regulation of yCT Expression yGT is usually considered a key component in cellular antioxidant defense since it is central to the recovery of glutathione; it is the only enzyme capable of initiating the degradation of this tripeptide for the subsequent recovery of the cysteine/cystine and resynthesis of GSH. However, under unique metabolic conditions, yGT can also play a pro-oxidant role [12]. yGT is a type II membrane protein with an uncleaved signal anchor (transmembrane domain,TM; see Figure 1) and a very short cytoplasmic domain consisting of only four polar amino acids. It is synthesized as a single chain propeptide and cleaved to yield a heterodimer with a large amphipathic subunit (50-60 kDa) and a small hydrophilic subunit (25-30 kDa); the size of the subunits on SDS-gels varies with celldependent differences in glycosylation. The catalytic active site includes residues from both subunits (blue triangles in Figure 1; reviewed in [13]). Previous studies from many groups have shown that this protein product is derived from a highly regulated gene in several different species (reviewed in [14]). The regulation involves alternative splicing coupled with alternative promoter usage that generates unique noncoding 5' ends for multiple yGT cDNAs. In contrast, splicing within coding

R.P. Hughey / y-Glutamyltranspeptidase Folding and Activation

148

exons appeared to be constitutive so that only a single protein product was encoded. Active site residues

N-glycan consensus sites:

A1

™

Tr^

N NN

N

NN

N IDD

'^

A2-5

A7 A8-9 Cysteine residues

4

+ 14»

in51Hi_i ' j! 3 4

1

+8aa

6't4 5

r>iA««««

j

IA

••A

"1rcf

CXXXC Fig. 1. Linear model of fGT protein structure.

In a collaboration with Drs. Martin Joyce-Brady and Jyh-Chang Jean at The Pulmonary Center of Boston University School of Medicine, we have now shown that alternative splicing events in mouse yGT also alter coding exons to produce at least three additional yGT-related protein isoforms (see Figure 1) [15]. Drs. Joyce-Brady and Jean first noted these alternative splicing events while characterizing the mutation that blocked yGT expression in the GGT**1 mouse [16,17]. They subsequently found that alternative splicing within exons 2 and 5 deletes residues 96-230 from the large subunit domain of the propeptide (yGTA2-5); alternative splicing within intron 7 introduces a premature stop codon in the large subunit after addition of 14 unique residues (yGTA7); and alternative splicing between exons 8 and 9 deletes residues 401-444 and introduces 8 new amino acids in the small subunit domain (yGTA8-9). Most importantly, the yGTA7 and A8-9 splicing events are developmentally regulated in heart, lung and kidney, and that for yGTA7 is shared with human yGT. Therefore, we believe that these alternative splicing events are physiologically relevant, and that these yGT protein isoforms serve unique roles in glutathione metabolism.

5. Expression and Characterization of yCT Isoforms [15] She directed mutagenesis of the mouse yGT(wt) cDNA (kindly provided by Dr. Michael W. Lieberman at Baylor College of Medicine in Houston) was used to prepare cDNAs with coding sequences corresponding to the yGTAl, A2-5, A7 and A8-9 transcripts in the expression vector pCDNA3 (Invitrogen). yGTAl is identical to yGT(wt) within the coding sequence, but exhibits a CAG insert upstream of the start site due to alternative splicing. Chinese Hamster Ovary Cells (CHO, nonpolarized epithelial cells) were transfected with the vectors, and stable clones were selected after growth in G418. Isoforms were also characterized after transient expression in CHO cells by transfection with plasmids (with T7 promoters) after the cells were infected with a

149

R.P. Hughey / y-Glutamyltranspeptidase Folding and Activation

recombinant cow pox virus (vT7CP) that overexpresses T7 RNA polymerase [18]. Whereas expression of yGTAl in CHO cells produces a 50-200 fold increase in yGT enzymatic activity in cell extracts, expression of the three new protein isoforms in CHO cells did not raise yGT activity above control levels in CHO cells (see Summary Table 1). Pulse-chase studies with [35S]Met/Cys, in conjunction with endoglycosidase H (endo H) treatment and cell surface biotinylation, were carried out with both clonal CHO cell lines and transiently transfected CHO cells. Treatment of immunoprecipitates with endo H removes N-glycans from glycoproteins that have not reached the processing enzymes of the medial Golgi complex in the biosynthetic pathway and provides a simple assay to measure exit of proteins from the ER based on altered mobility on SDS-PAGE (~3 kDa per N-giycan). Treatment of cells with the membrane-impermeant sulfo-NHS-SS-biotin tags proteins on the plasma membrane and provides an assay for delivery of proteins to the cell surface by recovery of biotinylated proteins from yGT immunoprecipitates with avidin-conjugated beads.

1

SUMMARY TABLE 1

1 1

^TA<

YGTA2-5

vGTA7

vGTAS-9

PEPTIDE (kDa) +/- endo H Mr difference No chase

24

9

12

15

2 h chase

none

9

12

15

endo Hs

endo Hs

endo Hs)

ER

ER

ER

(endo H" Subcellular localization (surface biotinylation) 2 h chase

surface

(immunofluorescence) surface Enzyme activity? yes

ER

ER

ER

no

no

no

Half-life (h) [MSHsoform clones

19

2

7

transient

17

2

10

not available

1

The results of our experiments [15] indicate that mouse yGTAl (full-length) is synthesized as a glycopeptide of 78 kDa with 6 N-glycans (see summary Table 1). It is subsequently cleaved to yield two subunits that are endo H resistant after 2 h, and delivered to the cell surface. yGTA2-5 is synthesized as a glycopeptide of 47 kDa with 3 N-glycans that never become resistant to treatment with endo H; yGTA7 is synthesized as a glycopeptide of 44 kDa with 4 N-glycans that remain endo H sensitive; and yGTA8-9 is synthesized as a glycopeptide of 52 kDa with 5 N-glycans that remain endo H sensitive. The apparent MW of yGTA8-9 also indicates that the protein is cleaved such that a small peptide is released and presumably degraded. Since neither yGTA2-5, A7 nor A8-9 is found at the cell surface by biotinylation, it was likely that these three isoforms were retained in the ER. This was confirmed by indirect immunofluorescence microscopy (not shown).

150

R.P. Hughey / f-Glutamyltranspeptidase Folding and Activation

and retained in the ER for subsequent degradation; and the very short half-life determined for the [35S]yGTA2-5 (Ih) and [35S]yGTA8-9 (2 h) is consistent with this possibility (tn=17 h for [35S]yGTAl). However, the half-life of [35S]yGTA7 (7 h) is significantly longer than the other two truncated isoforms suggesting that localization of yGTA7 could reflect a new, previously undefined role in the ER. This possibility is supported by finding that stable cell lines expressing either the full-length yGTAl or yGTA7, but not cell lines expressing either yGTA2-5, A8-9 or control glycoproteins (hCAR or hMUCl) mediate an ER stress response when the media was switched from DMEM/Ham's F12 (cysteine and cystine) to DMEM (cystine) only. Northern blot analysis of RNA from these cells showed a dramatic increase in the level of BiP expression and a clear induction of CHOP-10 only in cells expressing yGTAl or A7 (see Figure 2). Induction of BiP and CHOP-10 is a hallmark of the unfolded protein response (UPR) (reviewed in [19])[20,21].

Fig. 2. T<3TA1 and yGTAT mediate an ER stress response in CHO cete. Stable eel ines expressing yGT or control proteins were moved from DME+F12 (Cys ox + red) to DME (Cys ox) meola for 12 h and then returned to DME+F12. RNA was analyzed after 7 d recovery [14].

6. Characterization of yGT Bisulfide Bridges Present studies in my lab are now focused on identifying the features) of yGTAl and A7 overexpression that could initiate an ER stress response. This was addressed initially by aligning the yGTAl and A7 sequences (see Figure 1). In doing so, the most notable feature that is present in both isoforms is an unusual CX3C motif (LCEVFCR) formed by the fourth and fifth Cys in the sequence (numbering from the N-terminus). This motif is perfectly conserved in mouse, rat, pig and human yGT. Closely spaced protein drthiols are present in a variety of proteins with diverse functions. Thioredoxin reductase has a conserved CVNVGC sequence (CX4C) in its redox catalytic she, while thioredoxin (WCGPCK) and protein disulphide isomerase (WCGHCK have conserved CX2C motifs in their catalytic sites [22,23,24]. Finally, CX3C sequences are both a signature motif for one class of chemokines [25] and an essential coordinate for copper binding in several yeast proteins [26].

R.P. Hughey / Y-Glutamyltranspeptidase Folding and Activation

151

Fig. 3. Transient espression and characterization of yGT and Cys-mutants in CHAD cells. CHO cells were infected with vT7CP and transfected with cDNAs for either wt yGT or mutants with one or two Cys changed to Ala. Numbers refer to Cys residues in Fig. 1. Steady state levels of yGT were assessed by either Western blot analysis (A), or by labeling overnight with [^SJMet/eys (B). Surface expression of [^S] yGT (C) was determined by biotinylation of intact cells after a 30 min pulse and 2 h chase. See text for discussion.

The presence of disulfide bridges in yGT has not been previously addressed although all six Cys are highly conserved in mammalian and yeast yGT. The two subunits formed by yGT propeptide cleavage are clearly not linked by disulfides since the subunits readily separate on SDSPAGE in the absence of reducing agents. Since there is only one Cys (C6) in the small subunit, this means that C6 does not form a disulfide bridge. Since one Cys is within the transmembrane domain (Cl) and there is no evidence for dimerization of yGT propeptide or large subunit, Cl is also unlikely to form a disulfide bridge. However, disulfides could form between the other four residues (C2, C3, C4 and C5). Since mutagenesis of Cys residues that normally form disulfide bridges can cause ER retention and subsequent degradation, we used this approach to identify any disulfide bridges present in yGT. We mutated each Cys residue in yGT (wt, rat full-length cDNA)

152

R.P. Hughey / y-Glutamyltranspeptidase Folding and Activation

individually to Ala and the mutants were transiently expressed in CHO cells. The day after transfection, cells were extracted with detergent and yGT was immunoprecipitated for Western blot analysis (see Figure 3 A). While the native yGT(wt) reveals predominately propeptide cleaved to heterodimer (L+S) with extensively processed glycans (with characteristic diffuse bands) on P, L and S, this was clearly lacking for three mutants (C2, C3 and C5). Biotinylation of the ceU surface after metabolic labeling of the same transfectants showed a similar pattern although low levels of [3SS]C5 were clearly present (see Figure 3C). Surprisingly, mutant C4 shows almost normal levels of expression. However, closer inspection of the sequence around C4 also shows an adjacent glutamate residue (LC4E) and previous reports indicate that an adjacent negative charge can often rescue ER exit of proteins with a critical Cys mutation. Since mutation of both Cys residues from the same disulfide bridge can sometimes rescue protein expression [27], additional yGT mutants were prepared with all combinations between C2, C3, C4 and C5. Similar analysis of these mutants by Western blotting and biotinylation of [35S]-labeled cells revealed normal processing only for the double mutant of C4 and C5 (named 4,5; see Figure 3 A and 3C). While our cumulative results made it very clear that a critical disulfide bridge was formed between C2 and C3, we were not convinced that a disulfide bridge was present between C4 and C5. To determine if a disulfide bridge was present between C4 and C5, we tested the stability of the proteins both in vivo and in vitro since the absence of a stabilizing disulfide bridge could enhance denaturation and turnover of yGT. The in vivo stability of the yGT mutants was addressed by estimating the protein half-life for yGT(wt) and mutants that showed sufficient levels of expression by labeling with [3SS]Met/Cys. Transiently transfected cells expressing yGT(wt) or mutants Cl, C5, C6 or double mutant C4,5 were pulse labeled for 30 min and chased for either 3 h or 24 h. Imunoprecipitated [35S]yGT was quantified after SDS-PAGE using a phosphorimager (BioRad), and the half-life calculated from the decrease in [35S]yGT over 17 h was found to be no different between yGT(wt) and all four mutants (tH ~ 20 h; data not shown). In vitro stability of yGT(wt) and mutants was assessed by measuring the thermal denaturation properties of enzymatically active yGTs in Triton X-100 cell extracts. Cell extracts expressing yGT activity above control levels in CHO cells included yGT(wt), Cl, C4, C5, C6 and double mutant C4,5. Preliminary experiments indicated that heating the yGT(wt) extract for 15 min at 60°C prior to assay destroyed approximately half of the enzymatic activity subsequently measured at room temperature using y-ghitamyl-p-nitroanilide and glycylglycine as substrates [28]. Similar treatment of extracts from Cl- and C6-transfected cells showed thermal denaturation profiles no different than yGT(wt) (tw -10 min) as expected (see Figure 4). However, yGT activity in extracts from cells expressing mutants C4, C5 or double mutant C4,5 was significantly more sensitive to thermal denaturation (t*~5 min) consistent with the presence of a stabilizing disulfide bridge in yGT(wt) between C4 and C5.

7. yGT Enzymatic Activity Correlates with Propeptide

Cleavage

In the course of characterizing the expression of yGT with Cys mutations we noted that the mutants that remained uncleaved (C2, C3, and all double mutants except C4,5) also did not express yGT enzymatic activity above control levels in CHO cells. Since it is difficult to quantify levels of propeptide and heterodimer subunhs from a Western blot, transiently transfected cells were labeled overnight in [35S]Met/Cys to approximate steady state levels of yGT. In fact the overall pattern of

R.P. Hughey / -y-Glutamyltranspeptidase Folding and Activation

153

[35S]yGT (wt and mutants) obtained by this approach is quite similar to the pattern observed by Western blot analysis of the cells (compare Figures 3 A and 3B). Duplicate cell extracts were assayed for yGT enzymatic activity and a qualitative comparison of the profiles shows a dramatic correlation between levels of activity and heterodimer (L+S, but not P+L+S) indicating that the uncleaved propeptide (P) is not active (see Figure 5). Since the uncleaved mutants lacking Cys residues could represent misfolded proteins due to the absence of disulfide bndges we also prepared a mutant of yGT that changed Thr380 to Asn (T/N) within the highly conserved cleavage (*) site for heterodimer formation (DDG/AG * T380AHL). As expected, transient expression of our T380N mutant in CHO cells produced only the propeptide form of yGT by metabolic labeling with [35S]Met/Cys (30 min pulse and 24 h chase) and cell extracts did not exhibit yGT activity above control levels in mock-transfected cells (see Figure 6A and 6B). While a Thr to Cys (T380C) mutation also blocked propeptide cleavage, a more conservative substitution with Ser (T380S) did produce a significant level of heterodimer within 24 h. Although the level of [35S]heterodimer from yGT(T380S) expression was comparable to the level of [35S]heterodimer from yGT(wt) expression, there was significantly less enzyme activity found in the T380S cell extract indicating that the yGT active site was dramatically altered by this mutation of Thr380.

Fig. 4. Heat stability of yGT and Cys-mutants. Cell extracts were prepared from cells transiently expressing wt yGT or Cys mutants and heated for 15 min at 60°C before assay of yGT enzymatic activity.

154

R.P. Hughey / y-Glulamyllranspeplidase Folding and Activation

vrf

1 2 3 4 9 6

232420343040

yGT activity

E

i

v
1 2 3 4 5 6

232425343545

Fig. 5. yGT activity correlates with propeptide cleavage. Dupfcate plates of eels described in Fig. 3B were assayed for enzyme activity and levels of propeptide (P) and heterodmer (L+S) were quantified using a phosphorimager.

A.

mock wt T/N T/C T/B

mock wt T/N T/C T/S

Fig. 6. Thr380 is required for both yGT propeptide cleavage and enzyme activity. CHO cete transiently expressing either wt-yGT or cleavage mutants with Thr380 (T) changed to either Asn (N). Cys (C) or Ser (S) were pulse labeled for 30 min with [^SJMet+Cys and chased 24 h (A) or assayed for iGT activity the following day (B).

155

R.P. Hughey / y-Glutamyhranspeptidase Folding and Activation

In bacterial yGT only the distal side of the propeptide cleavage site is conserved (*T391TH); and mutation of either His393 to Gly or mutation of the Thr391 to Ala in E. coli yGT completely blocks propeptide cleavage [29]. However, mutation of Thr391 to Ser, or mutation of the Thr392 to Ala or Ser only partially blocks the cleavage. Most importantly, Hashimoto et al. [29] showed that there was a clear correlation between the efficiency of yGT propeptide cleavage and levels of eiKymatic activity in the periplasmic fraction of each transformed bacterial strain. They also showed that the uncleaved propeptide is delivered to the periplasmic space along with the heterodimer. Since we have previously shown that the uncleaved propeptide of rat yGT is delivered to the cell surface along with the heterodimer [30], pulse-chase studies were carried out to determine if the yGT T380N mutant was normally processed and delivered to the cell surface. Transiently transfected CHO cells were pulse labeled for 30 min with [35S]Met/Cys and chased for up to 135 min before biotinylation of the cell surface. Analysis of the [35S]yGT bands with a phosphorimager indicates that the yGT(T380N) is delivered to the cell surface with kinetics similar to the yGT(wt) (see Figure 7). The calculated half-life of [35S]yGT(T380N) was also not significantly different than yGT(wt) (data not shown). The cumulative data indicates that the absence of propeptide cleavage does not alter the stability or trafficking of the protein.

Surface fAvgdin-pptl

Total MQ%1 Chase t-46 90 135

4690136

4690136

4690 136 min -97

-66 -46 S-

Fig. 7. yGT cleavage is not needed for normal membrane trafficking to the cell surface. CHO cells transiently expressing either wt yGT or the T380N cleavage mutant were pulse labeled for 30 min with t^SJMet/Cys and chased up to 135 min. At each time point the cells were treated with a membrane impermeant biotinylation reagent. Cell surface yGT or T380N was recovered from the immunoprecipitates using avidin-conjugated beads.

156

R.P. Hughey / y-Glutamyltranspeptidase Folding and Activation

+/- Endo H or BtoBn -

*

B

*

B

+ B

B

-97

PM

L-

-45

S-

Mock

7
T380N

CD74*yOT

Fig. 8. yGT can be cleaved when retained in the endoptasmfc rettcukmi (ER). CHO cefts expressing either wt yGT, the T380N cleavage mutant or a chimera of yGT with the cytoptesmic domain of an ER resident protein CD74 were pulse labeled with 30 min with jHSJMetfCys ™d chased for 4 h before either biotinyiation of the eel surface or treatment of immunopretipitates +A- endo H.

Our finding that the yGT(T380S) mutant shows limited levels of propeptide processing with less enzymatic activity than expected, is consistent with the recent proposal by Inoue et al. [31] that the bacterial yGT is a member of the N-terminal nucleophile (Ntn) hydrolase family. This structural superfamily first described by Brannigan et al [32] exhibits a four-layered catalytically active offtacore formed by two antiparallel 3-sheets packed against each other and overlaid with a layer of antiparallel a-helices on one side. All of the Ntn-hydrdases exhibit auto-catalytic cleavage where the new amino-terminus (Thr, Ser or Cys) becomes the catalytic nucleophile for the enzyme. Other than this unique structure within the active site, the proteins have no other features in common. Based on the crystal structure and known propeptide cleavage events, members of this class include penicillin G acylases, proteasomes, ghitamine 5-phophoribosyl-l-pyrophosphate amidotransferases (GAT), aspartylghicosaminidase and I^aminopeptidase-D-Ala-esterase/amidase (see [33] for review). Brannigan et al. [32] first suggested that yGT might be part of this Ntnhydrolase family because of (i) the known propeptide cleavage event, (ii) finding Thr as the new amino-terminus, and (iii) the ability of yGT to use Gin as a substrate like another member, GAT. Although only a preliminary crystal structure of yGT is available [34], Inoue etal [31] have now shown that the small subunit amino-terminal Thr391 can be labeled by a novel mechanism-based affinity reagent that is a y-phosphonic acid monofluoride derivative of glutamic acid. Since Thr391 is the only conserved Cys, Ser or Thr within all known yGTs, the cumulative data provides strong evidence that yGT is a member of the Ntn-hydrolase superfamily. If yGT is a member of the Ntn-hydrolase superfamily then it is likely that the propeptide

R.P. Hughey / y-Glutamyltranspeptidase Folding and Activation

157

autocatalytic cleavage can occur as soon as the protein is properly folded within the endoplasmic reticulum (ER). Since the ER contains relatively high concentrations of glutathione it is hard to rationalize how active yGT could be present within the same compartment. One possibility is that YGT is associated with a specific inhibitory-chaperone within this compartment. In this case, overexpression of yGT(wt) by stable transfection could increase steady state levels of yGT in the ER which would bind all of the inhibitory-chaperone under oxidative stress condition and induce the unfolded protein response (UPR) as we observed for overexpression of yGTAl and A7. In fact, the UPR in yeast and mammals is known to be regulated by signaling through receptors in the ER/nuclear membrane when free levels of chaperones decrease due to increased levels of misfolded proteins (reviewed in [20]). Overexpression of the yGTA7 that is retained in the ER (but not control glycoproteins hCAR or hMUC) would interact with the same yGT-specific inhibitorychaperone under oxidative stress conditions and would explain our earlier observations. To increase the chance of isolating this yGT-specific chaperone we have prepared a chimera of yGT that should be retained in the ER. The chimera was made with the transmembrane and cytoplasmic domain of CD74 (Iip33), the MHC invariant chain isoform that is retained in the ER [35]. CD74 uses two alternative initiator Met residues that produces a type 2 membrane protein (N^VC0"*) with cytoplasmic domains of two different lengths, Iip31 and Iip33 The Iip33 has 16 additional residues including an amino-terminal Aig-rich sequence; at least two Arg are required for ER retention [35]. To determine if the chimera is actually retained in the ER, cells transiently transfected with yGT constructs were pulse labeled for 30 min and chased for 4 h to allow proteins to be fully processed and delivered to the cell surface. Whereas N-glycans on yGT(wt) and the T380N cleavage mutant were fully processed and resistant to cleavage by endo H, the chimera was fully sensitive to endo H treatment consistent with retention in the ER (see Figure 8). This was confirmed by biotinylation of the cell surface. While a third of the yGT(wt) and the T380N was recovered with avidinconjugated beads, less than 10% of the chimera was recovered in the biotinylated fraction. Assay of cell extracts also revealed that the chimera was fully active consistent with its cleavage to yield the heterodimer; and the ratio of activity to [35S]chimera was comparable to the ratio observed for yGT(wt) in the same experiment.

8. Conclusion Our previously published studies [15] indicate that yGT and the yGTA7 isoform can mediate an oxidative stress response within the endoplasmic reticulum. Experiments designed to characterize a unique CXXXC motif common to these two proteins, that might act as a redox sensor, indicates that there are two disulfide bridges within the large subunit of yGT. One disulfide is present between the second and third Cys from the N-terminus (C2-C3), and the other disulfide is present between the fourth and fifth Cys within the CXXXC sequence (C4-C5). Mutagenesis experiments indicate that the C2-C3 disulfide is required for propeptide cleavage, enzymatic activity and exit of the protein from the ER. In contrast, the removal of the C4-C5 disulfide has no effect on yGT activity, half-life, synthesis, processing (including cleavage) or delivery to the cell surface. However, the C4-C5 disulfide mutant is significantly more sensitive than the yGT (wt) to thermal denaturation. Our cumulative data shows a clear correlation between propeptide cleavage and enzyme activity and supports the hypothesis that yGT is a member of the Nm hydrolase superfamily [32,31]. Members of this family are characterized by autocatalytic cleavage to produce a new Nterminal residue that acts as the nucleophile within the enzyme active site [32,33]. Since we also

158

R.P. Hughey / •y-Glutamyltranspeptidase Folding and Activation

see propeptide cleavage of an enzymatically active yGT chimera designed for ER retention, it is likely that newly synthesized yGT associates with an inhibitory chaperone. Experiments are in progress to assess this possibility.

References 1. Baud, L. and ArdaBou, R., Involvement of reactive oxygen species in kidney damage. British Medfcal Bufetin 49 (1999) 621-629. 2. Master, A., Glutathione deficiency produced by inhibition of its synthesis and its reversal; appfcation in research and therapy, Pharmac.Ther. 51 (1999) 155-194. 3. Curthoys, N. P., Renal handing of glutathione In: Vina, J., Glutathione: metabolism and physiological functions. CRC Press.Boston, 1990, pp. 217 -225. 4. Sies, H., Glutathione and its rote in cellular functions, Free Rad.Biol.Med. 27 (1999) 916-921. 5. GHbert, H. F., Protein ofeulfjde isomerase and assisted protein folding, Journal of Biological Chemistry 272 (1997) 29399-29402. 6. Hwang, C., Sinskey, A. J., and Loofeh, H. F., Oxidized redox state of glutathione in the endoplasmic retcuhim, Science 257(1999)1496-1501. 7. Frand, A. R. and Kaiser, C. A, The ERO1 gene of yeast is required for oxidation of protein dtthiols in the endoplasmic reticulum, Mot.Cefl 1 (1998) 161-170. 8. Cuozzo, J. W. and Kaiser, C. A., Competition between glutathione and protein thiote for disulphide-bond formation. Nature Cel Biology 1 (1999) 130-135. 9. Frand, A. R., Cuozzo, J. W., and Kaiser, C. A, Pathways for protein dsulphide bond formation, Trends Cel Biol. 10 (2000)203-210. 10. Scott, R. D., Hughey, R., and Curthoys, N. P., Rote of apical and basoteteral secretion in turnover of glutathione in LLC-PK1 eels, American Journal of Physiology 265 (1993) F723-F728. 11. CareM, S., Ceriotti, A., Cabibbo, A., Gassina, G., Ruvo, M., and Site. R., Cysteine and glutathione secretion in response to protein cfisulfide bond formation in the ER, Science 277 (1997) 1681-1684. 12. Cutrin, J. C., Zingaro, B., Camandcte, S., Boveris, A., Pompete. A., and Pol, G., Contribution of y-glutamyl transpeptidase to oxidation damage of ischemic rat kidney, Kidney Int 57 (2000) 526-533. 13. Taniguchi, N, y-Ghitamyl transpeptidase catalytic mechanism and gene expression, Adv.Enz.ReLMotec.BioL 72 (1998) 239-278. 14. Chikhi, N., Hofic, N., GueHaen, G., and Laperche, Y., y-Glutamyl transpeptidase gene organization and expression : a comparative analysis in rat, mouse, pig and human species, Comp.Biochem Physiol 122 (1999) 367-380. 15. Joyce-Brady, M., Jean, J.-C., and Hughey, R. P, Gamma-gkJtamyKransferase and its isoform mediate an endoplasmic reticulum stress response, Journal of Biological Chemistry 276 (2001) 9468-9477. 16. Jean, J. C., Harding, C. O., Oakes, S. M., Yu, Q., Held, P. K.. and Joyce-Brady, M., Gamma-gkitamvl transferase (GGT) deficiency in the GGT*1*11 mouse results from a single point mutation that toads to a stop codon in the first coding exon of GGT mRNA, Mutagenesis 13 (1998) 101-106. 17. Harding, C. O., Williams. P., Wagner. E , Chang, D. S., Wild, K, CoJwei, R. E , and Wolff, J. A, Mice with genetic gamma-glutamyl transpeptidase deficiency exhibit glutathionuria, severe growth faiure, reduced ffe spans, and infertility, Journal of Biological Chemistry 272 (1997) 12560-12567. 18. Ramsey-Ewing, A. and Moss, B., Recombinant protein synthesis in Chinese hamster ovary eels using a vaccinia virus/bacteriophage T7 hybrid expression system, Journal of Biological Chemistry 271 (1996) 16962-16966. 19. Sidrauski, C., Chapman, R., and Walter, P., The unfolded protein response: an intracelular signafng pathway with many surprising features, Trends Cel Biol. 8 (1998) 245-249. 20. Spear, E. and Ng, T. W., The unfolded protein response: no longer just a special teams player, Traffic 2 (2001) 515523. 21. Ma, Y. and Hendershot, L M., The unfolding tate of the unfolded protein response, Cel 107 (2001) 827-830. 22. Mustacich, D. and Powis, G, Thioredoxin reductase, Biochem.J. 346 (2000) 1-8. 23. Powis, G. and Montfort, W R., Properties and biological activities of thioredoxins, Annu Rev Biophys Biomol Struct 30(2001)412-455. 24. Edman. J. C., EMs, L, Blacher, R. W., Roth, R. A, and Rutter, W. J., Sequence of protein ofeulphide isomerase and implication of its relationship to thioredoxin, Nature 317 (1985) 267-270. 25. Bazan, J. F., Bacon, D. B., Hardiman, G., Wang, W., Soo, K., Rossi, D.. Greaves, 0. R., Zkrtnik, S., and Schal, T. J., A new class of membrane-bound chemokine with a CXaC motif, Nature 385 (1997) 640-644. 26. Nittis, T., George, G. N., and Winge, D. R., Yeast Sco1, a protein essential for cytochrome c oxidase function is a Cu(l)-binding protein, Journal of Biological Chemistry 276 (2001) 42520-42526. 27. Darling, R. J., Ruddon, R. W., Perini, F., and Bedows, E., Cystine knot mutation s affect the folding of the gtycoprotain hormone a-subunit. Journal of Biological Chemistry 275 (2000) 15413-15421. 28. Hughey, R. P. and Curthoys, N. P., Comparison of the size and physical properties of r-glutarnyttransDepidase purified

R.P. Hughey / -y-Glutamyltranspeptidase Folding and Activation

\ 59

from rat kidney following solubilization with papain or with Triton X-100, Journal of Biological Chemistry 251 (1976) 7863-7870. 29. Hashimoto, W., Suzuki, H., Yamamoto, K., and Kumagai, H., Effect of site-directed mutations on processing and activity of y-glutamyttranspeptidase of Escherichia coli K-12, J.Biochem. 118 (1995) 75-80. 30. Altman, R. A. and Hughey, R. P., The identification of two subceflular sites for cleavage of y-glutamyltranspeptidase propeptide, Btochem.lnt. 13 (1986) 1009-1017. 31. Inoue, M., Hiratake, J., Suzuki, H., Kumagai, H., and Sakata, K., Identification of catalytic nudeophie of Escherichia coli y-glutamyltranspeptidase by y-monofluorophosphoono derivative of glutamic acid: N-terminal Thr-391 in smalt subunit is the nudeophile, Biochemistry 39 (2000) 7764-7771. 32. Brannigan, J. S., Dodson, G., Duggteby, H. J., Moody, P. C. E., Smith, J. L, Tomchick, D. R., and Murzin, A G., A protein catalytic framework with an N-terminal nucleophite is capable of self-activation, Nature 378 (1995) 416-419. 33. Oinonen, C. and Rouvinen, J., Strutural comparison of Ntn-hydrolases, Protein Science 9 (2000) 2329-2337. 34. Sakai, H., Sakabe, N., Sasaki, K., Hashimoto, W., Suzuki, H., Tachi, H., Kumagai, H., and Sakabe, K., A preliminary description of the crystal structure of y-glutamyltranspeptidase from E Co// K-12, J.Biochem. 120 (1996) 26-28. 35. Schutze, M. P., Peterson, P. A, and Jackson, M. R, An N-terminal doubie-arginine motif maintains type II membrane proteins in the endoplasmic reticulum, EMBO Journal 13 (1994) 1696-1705.

160

Thiol Metabolism and Redox Regulation of Cellular Functions A. Pompella et al. (Eds.) IOS Press, 2002

The Role of y-Glutamyl Transpeptidase in the Biosynthesis of Glutathione Avishay-Abraham STARK1, Noga PORAT1, Gloria VOLOHONSKY1-2, Arthur KOMLOSH1, Evgenia BLUVSHTEIN1, Chen TUBI1 and Pablo STEINBERG3 1 Dept. of Biochemistry, Tel Aviv University, Ramat Aviv 69978, Israel; 2 Present address: Dept. of Molecular Genetics, Weizmann Institute, Rehovot 76100 Israel, and 3 Inst. fur Erndhrungtoxikologie, Universitat Potsdam, Atrhur Scheunert-Allee 114-116, D14588 Bergholtz-Rehbriicke, Germany 1. The Ubiquity of y-GIutamyl Transpeptidase in Carcinomas and its Role in Carcinogenesis: Current Models and Unsolved Questions Y-Glutamyl transpeptidase (GOT), a glutathione (GSH) catabolizing, plasma membrane enzyme, is induced to high levels in many lesions at early stages of hepatocarcinogenesis (HC). The correlation among GGT-rich altered hepatic foci (AHF) in early/mid stages of HC, and the incidence and enzymatic profile of the carcinomas in later stages is the basis for the definition of foci and nodules as preneoplastic. Increased GGT levels in rodent and many human hepatic and other carcinomas, indicate that GGT may provide an advantage to cells of epithelial origin during carcinogenesis.

2. GGT and Resistance to Acute Toxicity This advantage has been attributed to participation of GGT in the detoxification of carcinogens, thus rendering cells more resistant to acute toxicity or to mitoinhibition. Resistance may not be a major contributor to the advantage because GGT-rich foci developed in animals treated with low, subacute doses of carcinogens, and transfection of cells with SV40, N-ras, or rasT24 led to increased GGT expression and tumorigenicity.

3. Mutagenesis and Lipid Peroxidation by the GSH-GGT System The first report on the activation of GSH to a mutagen by the GGT-rich kidney microsomes, but not by the GGT-poor liver microsomes [1] led us to test whether GGT was responsible for this activation. This turned out to be the case [2]. We thus hypothesized that GGT could provide an advantage by catabolism of extracellular GSH leading to the induction of mutations, the accumulation of which is crucial for carcinogenesis. We have studied extensively the mechanism of mutagenesis by the GSH-GGT system and have shown the following: Mutagenesis depends on the reduced state of the thiol group, on molecular oxygen and on transition metals [3]. Cleavage of a y-glutamyl moiety from GSH by

A.-A. Stark et al. / The Biosynthesis of Glutathione

161

GOT forms cysteinylglycine. Thiols undergo redox reaction at the thiolate anion level. Thus at physiological pH, cysteinyiglycine (pKasn = 6.3) is readily oxidized. Mutagenicity of thiols and of other reduced sulfur compounds is correlated with their pKasH or their reactivity as reductants [4,5,6]. Thiol oxidation leads to the reduction of iron which, in turn, interacts with oxygen to form superoxide and HbC^. The features of GSH-GGT-dependent lipid peroxidation (LPO) are similar to those of GSH mutagenesis [7,8]. Superoxide and H2O2 are the penultimate mutagens in that superoxide dismutase, which forms H2O2 increases mutagenesis whereas catalase/peroxidase that catabolize H2C>2 and radical scavengers inhibit mutagenesis [6,7,8,9,10,11]. The ultimate mutagen seems to be a hydroxyl radical, which is formed during thiol autoxidation [12]. Lipids in microsomal membranes and free linolenic acid are readily oxidized by the GSH-GGT system. The reaction depends on transition metals, and requires chelated iron. Ferric iron is readily reduced by the GSH-GGT system. And the rate of LPO depends on the steady state of Fe2+. Catalytic amounts of copper promote GSH-GGT-Fe-dependent LPO by increasing the steady state of Fe2+ [13]. Such reactions could occur in vivo in that the biological metal carriers transferrin and ceruloplasmin promote GSH-GGT-dependent mutagenesis and LPO [8,10,11]. We have shown that during experimental hepatocarcinogenesis in rats, oxidative damage is localized in GGT-rich preneoplastic hepatic foci, nodules and carcinomas [14]. Despite of the seemingly convincing arguments for oxidative mutagenesis initiated by the GSH-GGT system as responsible for the growth advantage of GGT-rich preneoplastic cells, the following findings indicate that the growth advantage may reside elsewhere, (i) There is no evidence for the induction of mutagenesis or oxidative damage to cellular DNA exerted by the GSH-GGT system in cultured mammalian cells (AA Stark, M Wellman-Rousseau, unpublished results), (ii) GGT-rich preneoplastic foci seem to be resistant to oxidative damage [15]. (in) Antioxidants that abolish GSH-mutagenesis and LPO are tumor promoters in the liver, skin and forestomach [16,17]. Thus, the advantage provided by GOT may result from other GGT-dependent reactions.

4. Development of Experimental Tools 4.1, Enzymatic assays for y-glutamylcysteine synthetase and glutathione synthetase The determination of y-GCS and GS are crucial for the understanding of GSH metabolism. Current assays require expensive equipment or radiochemicals. We developed spectrophotometric assays for y-GCS and GS. The assays are performed with crude extracts of cultured cells and tissues, and represent a novel combination of two known methods. y-GCS assay is based on the formation of GSH from cysteine, glutamate and glycine in the presence of purified rat kidney GS. The assay of GS is based on the formation of GSH from y-GC and glycine. GSH is then quantified by the recycling Tietze method. Reactions include the GGT inhibitor acivicin in order to prevent the degradation of y-GC and of the accumulating GSH, and DTT in order to prevent the oxidation of cysteine and y-glutamylcysteine. Multiple assays can be readily performed in a microplate reader [18].

162

A.-A. Stark el al. /The Biosynthesis ofGlutathione

4.2. Monoclonal antibodies that inhibit GGT In order to distinguish between the supply of cysteine that depends on by hydrolysis by GGT, and that of yGC that depends on transpeptidation by GGT, we have developed monoclonal antibodies that inhibit specifically the hydrolytic or the transpeptidatic activity of GGT. GGT bound to ELISA plates loses its enzymatic activity due to denaturation whereas plate-bound rat kidney brush border membrane vesicles retain GGT activity [19]. Hybridomas were prepared from mice immunized with purified, native GGT. Screening with membranal GGT resulted in the isolation of monoclonals that inhibit rat, but not mouse or human GGT. The Ab's recognize protein rather than against sugar epitopes in that each reacted with all GGT isoforms. Extensive kinetic analyses of free GGT and GGT-Ab complexes with D-Y-glutamyl-p-nitroanilide as substrate in the presence or the absence of maleate, or in the presence or absence of alanine, cysteine, cystine and glycylglycine as y-glutamyl acceptors revealed that Ab's 2A10 and 2E9 inhibited the hydrolytic and glutaminase activities and had little effect on the transpeptidation, whereas Ab's 4D7 and 5F10 inhibited transpeptidation, but not hydrolysis or glutaminase activities. Ab 5F10 mimicked the effect of maleate on GGT, in that it inhibited transpeptidation, enhanced the glutaminase activity and increased the affinity of the donor site of GGT to acivicin [20].

5. GGT supplies Cells with Precursors for GSH Biosynthesis GGT plays a key role in the transport of GSH constituents, leading to increase in cellular GSH. The latter is required for proliferation and resistance [21,22,23,24]. GSH biosynthesis is performed by Y-gmtamylcysteine synthetase (yGCS) and GSH synthetase (GS) (Eq. 1,2). yGCS levels and the cellular pool of cysteine are rate limiting. yGCS is feedback inhibited by GSH, leading to a steady state of GSH. GGT catalyzes the hydrolysis of a Y-glutamyl bond (Eq. 3,4), and transpeptidation (transpeptidation, Eq. 5). It could therefore participate in the supply of GSH constituents by two mechanisms: (i) Removal (hydrolysis) of a y-glutamyl residue from exogenous GSH by GGT followed by cleavage of cysteinylglycine (CG) by a dipeptidase or aminopeptidase [25] (Eq. 6) provides the rate limiting cysteine. If the supply of monomeric cysteine were the prime role of GGT, then the repletion of cellular GSH and growth of GGT-proficient and GGT-deficient cells should not differ when the only source of cellular cysteine is exogenous cysteine. This does not seem to be the case: Although cells harboring an expressing GGT transgene recovered cysteine from the medium and grew faster when GSH was the exogenous cysteine source [26], and skin cells harboring an expressing GGT gene produced tumors with an average mass 3-fold larger than vector-transfected cells [16], mouse hepatoma cells transfected with GGT cDNA grew faster than untransfected cells at limiting cysteine concentrations [27]. GGT-deficient rat embryo fibroblasts which were irradiated and treated with 12-Otetradecaloyl phorbolacetate (TPA) that formed benign tumors, contained levels of yGC and of GSH lower than those in cells that were not treated with TPA. An additional irradiation step resulted in cells that expressed GGT, manifested high extents of acivicin-sensitive GSH repletion in a medium containing cysteine, and formed malignant tumors [28]. These examples may reflect another activity of GGT that is important for the repletion of GSH.

163

A.-A. Stark et al. /The Biosynthesis of Glutathione

(ii) Trcmspeptidation by GOT in the presence of a y-glutamyl donor (e.g. GSH, glutamine), and cyst(e)ine as acceptor yields Y-glutamylcyst(e)ine (yGC, Eq. 7). The latter is readily transported [29,30] and reduced to Y-glutamylcysteine, and is used directly by GS to form GSH. The rate limiting yGCS and its feedback inhibition by GSH are thus circumvented, allowing a new, higher steady state of cellular GSH (Scheme 1). 1. Glutamate + cysteine + ATP -» Y-glutamylcysteine + ADP + Pi (Y-glutamylcysteine synthetase) 2. Y-glutamylcysteine + glycine + ATP -» glutathione + ADP + Pi (glutathione synthetase) 3. R-NH-Y-glutamyl (e.g. GSH) + GGT -* y-glutamyl-GGT + R-NH2 (e.g. cysteinylglycine) 4. Y-glutamyl-GGT + H2O -* GGT + glutamate (hydrolysis) 5. Y-glutamyl-GGT + R'-NH2 (acceptor) -» R'-NH-Y-glutamyl + GGT (transpeptidation) 6. Cysteinylglycine -»cysteine + glycine (dipeptidase) 7. R-NH-Y-glutamyl (e.g., GSH, glutamine) + cyst(e)ine -»Y-g'utamylcyst(e)ine + R-NH2

y-glutamylcystine

cystienylglycine

OOH SH CH2

NH

transporter

GSH reductase 2 GSH GS-SGi

NH

^COOH

COOH

Y-glutamylcysteine

Scheme 1. Glutathione synthesis that depends on transpeptidation by GGT

^COOH

164

A.-A. Stark et al. / The Biosynthesis of Glutathione

6. The Growth Advantage of GGT-Rich Cells is Associated with TranspeptidationDependent Supply of y-Glutamylcysteine 6.1. Methodology In order to study the role of transpeptidation in the supply of GSH constituents, we measured the rate and extent of GSH repletion in GGT-proficient and GGT-deficient cells. Cell were grown to confluency in 30 mm diameter 6-well clusters. GSH depletion was achieved by starvation for cysteine (growth at 10 fiM cysteine overnight) or by exposure to 1 mM diethylmaleate for 1 h. The medium was replaced by Hank's balanced salt solution with glucose, and the wells were exposed to (if not stated othwerwise) 200 jiM thiol equivalents. Triplicate wells were used to determine the concentrations of GSH and of protein with time [25]. i i I i i i i i i [ i i r 1 With Ab5F10 Iwithout Antibody, 120 -Cell line. \iM cysteine

- Cell line, nM cysteine

100 CO

o 0)

jj/5 "0

O

48

96

144 0

48

144

Time, hours Fig. 1. The effect of Ab 5F10 on the growth of oval cell line*. After growth of OC/CDE22 eels (white symbols) and M22 cells (black symbols) at non limiting cysteine concentration (medium A) for 24 hrs, the medium was replaced by medium supplemented with the indicated concentrations of cysteine, without or with 100 fig/ml of Ab 5F10. Cells were counted at the indicated time points. Presented are means±S.D. of triplicate wells.

165

A.-A. Stark el at. /The Biosynthesis of Glutathione

6.2. Tumorigenic and non tumorigenic rat oval liver cell lines Non-tumorigenic rat liver oval cell line OC/CDE22, and its GGT-rich, tumorigenic counterpart line M22 were grown at limiting and non limiting cysteine concentrations in a glutaminecontaining medium. Save that of GOT, the levels of all enzymes involved in GSH biosynthesis and degradation were comparable (Table 1). The growth advantage of M22 is clear in that M22 grew faster than OC/CDE22. MAb 5F10 that specifically inhibits transpeptidation did not affect the growth rates of OC/CDE22, but decreased those of M22 to the OC/CDE22 level (Fig 1), indicating that transpeptidation (glutamine as donor and cysteine acceptor to form y-GC) appeared to be largely responible for the fast growth rates of M22. Interestingly, exposure to cystine increased levels of cellular GSH and facilitated the growth of the malignant hepatocellular carcinoma line HepG2 [31]. Table 1. Activity of enzymes involved in GSH biosynthesis and degradation

Cell line

OC/CDE22

M22

NIH3T3

S-13 (NIH3T3-GGThum)

GSH content, nmol/mg total protein

16±3(24)

34±5 (24)

20.4±1(10)

26±2 (10)

y-GC synthetase, mU/mg cytosolic protein

0.69±0.31 (24)

0.84±0.23 (24)

0.94±0.2 (3)

0.68±0.3 (3)

Glutathione synthetase, mU/mg cytosolic protein

1.66±0.31 (24)

4.6±0.44 (24)

2.1±0.18(3)

2.14±0.34(3)

2.3±0.3 (3) 0.4(1) ND ND

2.8±0.2 (3) 84±1 (3) ND 32(1)

1.28±0.49 ND(b) ND NT(C>

1.46±0.26 42±15(6) ND NT

MEMBRANES Aminopeptidase M(a) GGT*"' GGT with acivicin(a) GGTwithAb5F10(a> ATTACHED WHOLE CELLS(O) GGT<e> GGT with acivicin(e) GGT with Ab5F10(e)

ND ND ND

75 ±14 (10) 5.1±0.8(10) 30.4±7.1 (10)

NT NT NT

NT NT NT

Values of means±S.D. (number of determinations). Preincubation of cells with acivicin was at 0.5 mM for 60 min. Preincubation with Ab 5F10 was at 100 ng/ml for 45 min. (a) mU/mg membranal protein; (b) not detected; (c), not tested; (d) GSH-depleted cells, assayed in culture at 37°C. (e), mU/mg total protein. (Adapted from Ref 25).

In order to test whether transpeptidation leads to increased GSH content, GSH repletion was examined in cells exposed to GSH precursors that are (cyst(e)ine/glutamine) or are not (cyst(e)ine/glutamate) substrates for transpeptidation. The rate and extent of GSH repletion with cyst(e)ine/glutamine exceeded those obtained with cyst(e)ine/glutamate in M22, but not in OC/CDE22 cells. When transpeptidation was inhibited by Ab 5F10, repletion with cyst(e)ine/glutamine was similar to that obtained with cyst(e)ine/glutamate (Fig 2). Repletion

166

A. -A. Stark et al. I The Biosynthesis of Glutathione

with cystine/glutamine yielded the highest cellular GSH concentrations, probably because cystine is the most efficient y-glutamyl acceptor in transpeptidation [20,32].

1111111111111111111111111111n ~_ OC/CDE22

~

Cys. Glu Cystine, -

Cystine, Gin

Gin, 5F1Q lystine, Glu

Cystine. GU. Cys, Gin— Cys. Glu -

l 11 I 11 l 1 1 1 l l l I l l 11 11 1 1 1 1 1 l l l

I II I II I II III III I I III I

5

6

0

1

2

3

4

5

6

Time, hours Fig. 2. Repletion of GSH with cystine and glutamine In oval cells. GSH-depteted ceHs were exposed at time zero to HBSS containing 200 nM grycine and the indicated supplements at 200 jiM thiol equivalents, cysteine, glutamate, cystine or 2 mM glutamine. GSH and protein were determined at the indicated time points.

6.3. Transport of-y-glutamylcysteine supports GSH biosynthesis In the GGT-deficient OC/CDE22 cells, the extent of GSH repletion with y-GC was comparable to that obtained with cyst(e)ine, and acivicin did not inhibit repletion (Fig 3). BSO that inhibits y-GCS and the transport of y-glutamyl peptides [30] abolished repletion supported by y-GC (and by GSH and cyst(e)ine), indicating that y-GC was taken up as such by a putative y-GC transporter which is sensitive to BSO and resistant to acivicin, and was used for GSH synthesis. The putative transporter appears to have been rendered sensitive to acivicin in the heavily mutagenized [33] M22 cells.

A.-A. Stark et al. /The Biosynthesis of Glutathione

OC/CDE22

40

30

~O

E 20 c I Cfl 0

10

Time, hours Fig. 3. Repletion of GSH with y-glutamylcysteine In oval cells. GSH-depleted M22 and OC/CDE22 celts were exposed at time zero to HBSS containing, as indicated, 200 \M cysteine (•,€>) or y glutamylcysteine (V,V) in the absence (black symbols) or the presence (open symbols) of 0.5 mM acivicin for 60 min prior to exposure. GSH and protein were determined at the indicated time points.

60

o o. 40 CD

E "o E c

20

Cysgly, Bestatin Cys, BSO

0

i i I i i i

Time, hours Fig. 4. Effect of inhibitors of amlnopeptidase on GSH repletion, with cystelnylglycine as the cysteine source. GSH-depleted M22 cells were exposed at time zero to HBSS containing as indicated: 200 M.M GSH; 200 \M cysteinylglycine, 400 ^M penicillamine and 10 ^M bestatin. GSH and protein were determined at the indicated time points.

168

A.-A. Stark et at. / The Biosynthesis of Glutathione

6.4. GSH is not the preferred source for GSH precursors: cleavage of cysteinylglycine as the rate limiting step in the utilization of exogenous GSH Cysteinylglycine (CG) and cysteine lead to similar rates of GSH synthesis in the cysteinylglycinase-rich rat kidney. Oval cells lack cysteinylglycinase but contain aminopeptidase (Table 1). In GGT-proficient M22 (but not in the GGT-deficient OC/CDE22) cells, repletion with GSH was slow, and was abolished by acivicin (see below). Repletion with GSH or CG required aminopeptidase activity hi that inhibition of the latter by bestatin abolished completely repletion with GSH and CG, but not with cysteine, and was lower than that obtained with cyst(e)ine (Fig 4). The findings indicate that aminopeptidase/dipeptidase activity is required and rate limiting for GSH repletion when GSH or CG serve as cysteine sources. 6.5. NIH3T3 cells appear to have a y-GC transporter In order to test whether such transport of y-GC occurs in other cell lines, we used mouse fibroblast NIH3T3 cells. NIH3T3 are GGT-deficient. We have constructed cell line S-13, an NIH cells line that harbors the human GGT gene regulated by an inducible metallothionein promoter. Similarly to oval cells, line S-13, but not NIH3T3, could utilize exogenous GSH for repletion. Both lines manifested acivicin-resistant repletion with cysteine (data not shown). In NIH3T3, repletion rates with cystine/glutamate cystme/glutamine were comparable and resistant to acivicin. In S-13, the rate of repletion with cystine/glutamine was higher than that hi NIH3T3 without acivicin, whereas in its presence the rate of repletion was comparable to that in NIH3T3 with or without acivicin. Table 2. Rate of repletion of GSH with cysteine or Y-glutamylcy*tetne. See text for explanations. Cellular GSH content, nmd/mg protein

Cysteine Cysteine + acivicin

Y-GC Y-GC+ acivicin

9

20

15

13

11

24

14

10

1 1 5

18

40

21

22

35

18

18

i 1

9

14

5.5

9

9.5

4

5

7

5

18

22

14

18

c

16

4

12

15

i !

Y-GC Y-GC + acivicin

NIH3T3

OC/CDE22

18.5

Cysteine Cysteine + acivicin

S-13

M22

Time, hrs

i

5 1

Although the endpoint of repletion was the cellular concentration of GSH, the kinetics of repletion reflect the uptake of the rate limiting cysteine and its congeners rather than the kinetic properties of r glutamyteysteine synthetase and glutathione synthetase. The Km of the hotoenzyme (or the heavy subunit) of human and rat Y-GCS for cysteine is 196±96 \M [34], and that of glutathione synthetase is 50 nM [35]. The apparent Km for cysteine-supported repletion in NIH and oval cells was 25 nM, eightfold lower than that of Y-GCS, indicating that transport of cysteine was rate limiting. As judged by repletion rates, the apparent Km's for Y-GC in NIH3T3 and oval cells were 50 \M and 250 pM, respectively.

A.-A. Stark et al. / The Biosynthesis of Glutathione

169

Similarly to OC/CDE22, the repletion with y-GC in NIH3T3 was lower than that obtained with cysteine, and was resistant to acivicin. The extent of repletion with y-GC in S-13 was higher and approached that obtained with cysteine,, but was indistinguishable from that obtained with cysteine upon inactivation of GGT. This suggests that the higher repletion in S13 was due to the transport intact y-GC by transport and to the supply of cysteine originating in hydrolysis of y-GC by GGT. Inhibition of GGT activity unmasked the repletion supported by the transport of intact y-GC. Thus, NIH3T3and S-13 appear to have the putative y-GC transporter. The properties of the latter differ from those of the transporter in oval cells. We have measured the kinetics of the repletion of GSH in the GGT-deficient lines OC/CDE22 and NIH3T3 over a wide concentration range of y-GC or cysteine. The extent of repletion at 1 and 5 h time point were plotted against the concentration of y-GC and cysteine (Table 2). The fact that GSH efflux is decreased or inhibited upon treatment of cells with acivicin [36,37] has led to the notion that GGT participates in the transport of GSH or of its constituents. This work and the work of others [32, 32a] indicate that GGT functions in the biosynthesis, but not the transport of y-GC. Increased cellular GSH content has been frequently associated with increased GGT levels in cancer cells and tumors. Just as frequent are the reports about normal GSH content in GGT-rich cells. The findings presented here indicate that the difference may reside in the presence or absence of a transporter for y-GC, or in the properties of such a transporter. The yGC transporter is not one of the GSH transporters in that the latter inhibited by GSHderivatives, are resistant to BSO (and acivicin) and are not inhibited competitively by yglutamyl peptides [38,39,40]. It is unclear as of yet whether the transporter of y-GC is one of the known peptide transporter. Acknowledgments We wish to acknowledge the generous donations from the Jac and Eva Feinberg Foundation, New York and from Mr. Freddy Furhmann and Mrs. Ina Curiel, Curagao. Supported in part by the Ella Kodesz Institute for Research on Cancer Development and Prevention, the Tel-Aviv University Cancer Biology Research Center, and by the Rekanati Foundation for Medical Research Israel.

References 1

Glatt H, Protic-SabljiC M, Oesch F. Mutagenicity of glutathione and cysteine in the Ames test. Science 220:961963, 1983. 2 Stark, AA, Zeiger, E, Pagano, DA. Glutathione mutagenesis in Salmonella typhimurium TA100: dependence on a single enzyme, y-glutamyl transpeptidase. Mutation Res. 177:45-52,1987. 3 Stark AA, Zeiger E., Pagano DA. Glutathione mutagenesis in Salmonella typhimurium is a y-glutamyl transpeptidase-dependent process involving active oxygen species. Carcinogenesis 9:771-777,1988. 4 Stark AA, Pagano D.A., Zeiger E. Effect of pH on mutagenesis by thiols in Salmonella typhimurium TA102. Mutation Res. 224:89-94,1989. 5 Pagano DA, Zeiger E, Stark AA Autoxidation and mutagenicity of sodium bisulfite, Mutation Res. 228:89-96, 1990.

170

6

A. -A. Stark et al. / The Biosynthesis of Glutathione

Stark AA, Zeiger E, Pagano DA. Oxidative mutagenesis by the glutathione~y-glutamyl transpeptidase system: mechanism and relevance to hepatocarcinogenesis. In: Nygaard OF, Upton AC (eds.) Anticai&nogenesis and Radiation Protection, vol 2, Plenum Publishing Corp., New York, pp. 61-71,1991.

7

Stark AA. Oxidative metabolism of glutathione by Y-glutamyl transpeptidase and peroxisome proliferation: the

8

Stark AA, Zeiger E, Pagano DA. Glutathione metabolism by y-glutamyl transpeptidase leads to lipid

relevance to hepatocarcinogenesis. A hypothesis. Mutagenesis 6241-245,1991. peroxidation: characterization of the system and relevance to hepatocarcinogenesis. Carcinogenesis 14:183189, 1993. Stark AA, Pagano DA, Glass GA, Kamin-Belsky N, Zeiger E. The effects of antexidants and enzymes involved in glutathione metabolism on mutagenesis by glutathione and L-cysteine. Mutation Res. 308:215-222, 1994. 9

Stark AA, Glass GA. The role copper and cerutoplasmin in oxidative mutagenesis induced by the glutathione-rglutamyl transpeptidase system and by other thiols. Env. Mol. Mutagenesis 29:63-72,1997.

10 Glass GA, Stark AA. Promotion of glutathione-rglutamyl transpeptidase-dependent Hpid peroxidation by copper and cerutoplasmin: the requirement for iron and the effects of anttoxkJants and antioxidant enzymes. Env. Mol. Mutagenesis 29:73-80,1997. 11 Diez L, LJvertoux MH, Stark AA, Wellman-Rousseau M, Leroy P. High-performance liquid chromatographic assay of hydroxyl free radical using salicylic acid hydroxylation during hi vitro experiments involving thiols. J. Chromatogr B; Biomed Sci Appl. 763:185-93, 2001. 13 Zalit I, Glass GA, Stark AA. The role of transition metals in glutathione- yglutamyl transpeptidase-dependent lipid peroxidation: thiol-metal chelates as possible reactive intermediates. Biochem. Mol. Bid.

Interned.

40:1123-1133,1996 12 Stark, AA, Russell JJ, Langenbach RE Pagano DA, Zeiger E, Huberman E. Localization of oxidative damage by a glutathione-Y-glutamyl transpeptidase system in preneoplastic lesions in sections of livers from cacrcinogentreated rats. Carcinogenesis 15:343-348,1994. 15 16

Benedetti, A., MalvakJi, G., Fulceri, R. and Comporti, M. (1984) Loss of lipid peroxidation as a Nstcchemical marker for preneoplastic hepatocellular foci of rats. Cancer Res. 44, 5712-5717. Warren BS, Naytor MF, Winberg LD, Yoshimi N, Volpe JPG, Gimenez-Conti I, Slaga TJ. Induction and inhibition of tumor progression. Proc. Soc. ExptJ. Biol. Med. 202:9-15,1993.

17 Shirai T, Fujishima S, Ohshima M, Masuda A, Ito N. Effects of BHA, BHT and NaCI on gastric Carcinogenesis initiated with MNNG in F344 rats J. Natl. Cancer Inst. 72:189-1198,1984. 18 Volohonsky G, Tubi C, Porat N, Wellman-Rousseau M, Visvikis A, Leroy P, RasN S, Steinberg P, Stark AA. A Spectrophotometric Assay of y-Glutamylcysteine Synthetase and Glutathione Synthetase in Crude extracts from Tissues and Cultured Mammalian Cells. Chem Biol Interact. 2002, submitted. 19 Glass GA, Bluvshtein E, Gur Y, Sfez S, Simovich, V, Stark AA. Detection of conformational changes in rat kidney Y-glutamyl transpeptidase by an antibody against a synthetic peptide pertaining to a part of the reactive center of the enzyme. Biochem. Mol. Biol. Internet). 33:505-513,1994. 20 Bluvshtein E, Glass GA, Volohonsky G, Seidel A, Frank H. Steinberg P, Oesch F, Stark AA. Inhibition of the hydrolytic and transpeptidatic activities of rat kidney y-glutamyi transpeptidase by monoclonal antibodies. Eur. J. Biochem. 260:844-854, 1999. 21 Pitot HC. Altered hepatic foci: their rote in murine hepatocarcinogenesis. Ann. Rev. Toxicol. Phrmacol. 30:465-500, 1990. 22 Hanigan MH, Pitot HC Y-Glutamyl transpeptidase: its rote in hepatocarcinogenesis.

Carcinogenesis 6:65-172,

1985. 23

Cameron RG, Armstrong D, Gunsekara A, Varghses G, Speisky H. Utilization of circulating glutathione by nodular and cancerous intact rat liver. Carcinogenesis 12:2369-2372,1991. 24 Hendrich S, Pitot HC. Enzymes of glutathione metabolism as biochemical markers during hepatocarcinogenesis. Cancer Metast. Rev. 6:155-178,1987.

A.-A. Stark et al. / The Biosynthesis of Glutathione

171

25 Komtosh A, Volohonsky G, Porat N, Tuby C, Bluvshtein E, Steinberg P, Oesch F, Stark AA. Y-Glutamyl transpeptidase and glutathione biosynthesis in non tumorigenic and tumorigenic rat liver oval cell lines. Carcinogenesis 22:2009-2016, 2001. 26 Rajpert-De Meyts E, Shi M, Chang M, Robison TW, Groffen J, Heisterkamp N, Forman HJ. Transfection with Y-glutamyl transpeptidase enhances recovery from glutathione depletion using extracellular glutathione. Toxicol. Appl. Pharmacol. 114:56-62,1992. 27 Hanigan MH. Expression of y-glutamyl transpeptidase provides tumor cells with a selective growth advantage at physiologic concentrations of cyst(e)ine. Carcinogenesis 16:181-185,1995. 28 Sierra-Rivera E, Meredith MJ, Summar ML, Smith MD, Voorhees GJ, Stoffel CM, Freeman MLGenes regulating glutathione concentrations in X-ray-transformed rat embryo fibroblasts: changes in y-glutamylcysteine synthetase and y-glutamyttranspeptidase expression. Carcinogenesis 7:1301-1307,1994. 29 Griffith OW, Bridges RJ, Meister A. Formation of Y-glutamyteyst(e)ine in vivo is catalyzed by y-glutamyl transpeptidase. Proc. Natl. Acad. Sci. USA 78:2777-2781,1981. 30 Anderson ME, Meister A. Transport and utilization of f-glutamylcyst(e)ine for glutathione synthesis. Proc. Natl. Acad. Sci. USA 80:707-711,1983. 31 Huang Z-Z, Chen C, Zeng Z, Yang H, Oh J, Chen L, Lu SC. Mechanism and significance of increased glutathione level in human hepatocellular carcinoma and liver regeneration. FASEB J. 15:19-21, 2001. 32 Allison RD. y-Glutamyl transpeptidase: kinetics and mechanism. Methods in Enzymol. 113:419-437,1985. 33 Steinberg P, Steinbrecher R, Radaeva S, Schirmacher P, Dienes HP, Oesch F, Bannasch P. Oval cell lines OC/CDE-6 and OC/CDE-22 give rise to cholangiocellular and undifferentiated carcinomas after transformation. Lab. Invest., 71:700-709,1994. 34 Griffith, OW. Biologic and pharmacologic regulation of mammalian glutathione synthesis. Free Radical Biol. Med. 27:922-935, 1999. 35 Meister A. Glutathione synthetase from rat kidney Methods in Enzymol, 113:393-399,1985. 36 Griffith OW, Meister A. Translocation of intracellular glutathione to membrane bound Y-glutamyl transpeptidase as a discrete step in the Y-glutamyl cycle: glutathionuria after inhibition of transpeptidase. Proc. Natl. Acad. Sci. USA 76:268-272, 1979. 37 Griffith OW, Bridges RJ, Meister A. Transport of Y-glutamyl amino acids: role of glutathione and Y-glutamyl transpeptidase. Proc. Natl. Acad. Sci.USA 76:6319-6322,1979. 38 Adibi, SA. The oligopeptide transporter (Pept-1) in human intestine: biology and function. Gastroenterology 113:332-340, 1997. 39 Kaptowitz N, Fernandez-Checa CJ, Kannan R, Garcia-Ruiz C, Ookhtens M, Yi JR. GSH transporters: molecular characterization and role in GSH homeostasis. Biol. Chem. 377:267-273,1996. 40 Lu, SC, Sun, W-M, Yi J, Ookhtens M, Sze G, Kaptowitz N. Role of two recently cloned rat liver GSH transaporters in the ubiquitous transport of GSH in mammalian cells. J. Clin Invest. 97:1488-1496,1996.

172

Thiol Metabolism and Redox Regulation of Cellular Functions A. Pompella el al. (Eds.) 1OS Press, 2002

Role of y-Glutamyltransferase in the Homeostasis of Glutathione during Oxidative and Nitrosative Stress N.-E. HUSEBY1, N. ASARE1, S. WETTING1,1.M. MDCKELSEN1, B. MORTENSEN1 and M. WELLMAN2 1 Department of Medical Biochemistry, University of Tromse, Norway 2 Thiols etfonctions cellulaires, University Henri Poincare Nancy I, Nancy, France.

1. Introduction Glutathione (GSH) is an important and abundant antioxidant in most cells, including cancer cells. The GSH level can be significantly reduced during oxidative and nitrosative stress, and GSH depletion is an early step in apoptosis. As most cells have no uptake mechanism of GSH, they depend on resynthesis to prevent cell death. A limiting factor in cellular GSH synthesis is the availability of cyst(e)ine (for recent reviews, see [1,2]). During tumor progression cancer cells tend to develop invasive and metastatic growth and resistance to therapy. They also show metabolic alterations that are termed loss of differentiation, and frequently these changes include increased resistance towards oxidative and nitrosative stress [3,4]. It has been shown that GSH protects tumor cells against oxidative stress in liver microvasculature [5] and that tumor cells with elevated GSH or ability to resynthesise GSH possess a higher invasive potential [4,6]. NO is highly tumoricidal in the presence of H2O2 [6,7], and these agents appear to be part of a natural defence towards metastatic growth [8]. GSH plays an important role in the protection of tumor cells against NO mediated apoptosis [9,10] and intracellular GSH will therefore contribute to the mechanisms of tumor cell survival during metastatic growth. The enzyme y-glutamyltransferase (GGT) initiates the breakdown of extracellular glutathione by hydrolysing the y-glu-cys bond in GSH. The remaining part, cys-glu, is then hydrolysed by peptidases. Cellular uptake of the amino acids will thus provide the monomers needed for GSH biosynthesis [11] and it has been suggested that GGT has significant importance for GSH homoestasis when the level of cyst(e)ine is reduced [12-14]. Fibroblasts requiring cysteine can grow in cysteine-free medium if supplemented with GSH [12,15]. Cells cultivated in vitro have access to surplus amounts of cystine (200 uM) when compared to in vivo concentrations (5-10 uM). It was recently shown that incubation of lymphocytes in medium devoid of cystine resulted in oxidative stress and apoptosis, but lymphocytes transfected with GGT were protected if GSH was added to the medium [13]. These studies confirm that the activity of GGT resuhs in increased uptake of cysteine obtained by degradation of extracellular GSH and in this way protects the cells from GSH depletion and oxidation-induced cell death. We and other investigators have shown that GGT is induced after exposure of cancer cells to peroxides [16-20]. When this is linked to the ability of the enzyme to provide cyst(e)ine, it appears that GGT has an important role in GSH salvaging during oxidative stress. Our aim was to investigate whether the enzyme is also induced by nitrosative stress in a similar way, and

N. -E. Huseby et al. / Homeostasis of Glutathione

173

particularly whether GGT induction is associated with changes in intracellular GSH. We then set out to find whether GGT helps these tumor cells to avoid apoptosis in situations with depleted cysteine, by degrading GSH in the medium. We also investigated whether cells with induced GGT have an increased ability to obtain cysteine from such degradation. NO is a pleiotropic agent, and has been reported to deplete cells of GSH at higher concentrations resulting in cell death and apoptosis. We found that GGT was induced by NO, and that the enzyme activity reduced cell death when incubated in cyst(e)ine reduced medium.

2. Materials and Methods Cell lines and culture conditions CC531 is a colon carcinoma cell line that was originally developed hi rats after chemical carcinogenesis [21]. The cells were cultured in RPMI1640 medium with 5% fetal calf serum in a humid atmosphere with 5% COa at 37 °C. This cell line is metastatic to liver in syngeneic WagRij rats [19]. Treatment of cells for enzyme induction Viable cells were seeded in 6 well plates (0.2 x 106 cells per well) with 3 ml medium and allowed to proliferate overnight. They were then exposed to menadione (50 uM for 15 min) or spermine NONOate (SpNO, 200 uM for 4 h), followed by cultivation in normal growth medium for 2 days before being harvested. N-acetylcysteine (NAC) was added as an antioxidant to the medium during the oxidative or nitrosative stress incubation periods at a concentration of 10 mM. Cell harvesting and enzyme measurements Cells were harvested by trypsinization, solubilized at a concentration of 2 x 106 cells/ml in phosphate buffered saline (PBS) with 1% Triton X-100 and the GGT activity was measured [19]. To estimate the amount of dead cells floating in the medium after SpNO incubations, the medium was collected, centrifuged and the pellet was solubilized in 250 ul PBS with 1% Triton X-100 before measurement of lactate dehydrogenase (LDH). GGT and LDH activities were determined at 37 °C according to the recommendations of the International Federation of Clinical Chemistry (BFCC), with commercial kits (y-GT and LDH optimized kits, Boehringer Mannheim Lab Diagnostics, Germany). Measurement of reduced glutathione Cells (0.75 x 106) were seeded in 6 cm plates and incubated overnight in normal growth medium. After exposing the cells to meandione or SpNO, or incubating them in cystine-depleted medium, the cells were washed with cold PBS and harvested by scraping in 300 ul metaphosphoric acid (5%, w/v). The supernatant obtained after centrifugation at 12000 x g for 10 min at 4°C was analysed using a GSH assay kit (Calbiochem) which is based on the nonenzymatic reaction of thiols with a quinolinium chromogen. Crystalline GSH (Sigma-Aldrich) was used as standard. The precipitated proteins were solubilized in a volume of 300 ul 0.1 M NaOH and quantified. The GSH content was expressed as nmol GSH/mg protein.

174

N. -E. Huseby etal./ Homeoslasis of Glutathione

Apoptosis andDNA Fragmentation assays Quantification of apoptosis was detennined using the DNA Fragmentation ELISA kit (Boehringer Mannheim), according to the procedure recommended by the manufacturer. In brief, cells were seeded in 6-well plates (2 x 10s cells/well) and incubated for 24 h in control medium with 1 uM BrdU, then in cystine-depleted medium (with or without 25 uM GSH) for 24 h before being harvested for apoptosis assay. The final measurement was performed after stopping the ELISA reaction using H2SO4 and reading the absorbance at 450 nm. Data from experimental cells were shown relative to those of control cells. Uptake of radiolabelled cysteine Cells, either control cells or cells preincubated 96 h earlier with 250 uM SpNO for a 4 h period, were seeded in 6-well plates (2 x 10s cells/wen) with 3 ml normal growth medium. After 24 h, the medium was changed to cystine-depleted medium to which was added 35S-labelled GSH (purchased from Peridn-Elmer Life Sciences, Boston, USA) at 0.5 uCi/ml, 25 uM GSH and 25 mM glycylglycine. A series of control cells were also incubated in this medium with 500 uM acivicin. After incubation periodes of 15 min, 2 h, 4 h and 8 h, the cefls were washed twice with 1 ml PBS and solubilized in 1.0 ml 0.1 M NaOH with 0.1% SDS at 37°C. Radioactivity was counted after mixing 250 ul of cell lysate with 3.0 ml Ultima Gold in a liquid scintillation counter.

3. Results The effect ofmenadione and SpNO on cell proliferation and death The rat colon carcinoma cells were exposed to oxidative or nitrosative stress by incubations with 50 uM menadione for 15 min or 250 \iM SpNO for 4 h. These periods were followed by cultivation in normal growth medium for 24-48 h. Cell proliferation was significantly decreased, as shown by the reduced number of viable cells after 48 h (Fig. 1 A). At mis time, an increasing number of dead cells were detected floating in the medium. A quantitative estimate of cefl death was obtained by measuring the LDH activity after collecting the floating cells (Fig. IB). Increasing the concentration of either agent resulted in higher cell death. Oxidative and nitrosative stress induces GOT and temporarily consumes GSH in rat colon carcinoma cells To test whether oxidative or nitrosative stress induced GGT and concomitantly reduced GSH, the cells were incubated as stated above and harvested at various time points. Both menadione and SpNO exposures resulted in increased GGT activity (Fig. 2) which was detectable 48 h after initiation of the incubations. The increases were dose-related and could be blocked by adding antioxidants (Fig. 2 shows the effect of N-acetylcysteine). This blocking effect was almost total after menadione but only partial after SpNO incubations. The intracettular GSH level was significantly decreased when measured 4 h after the initiation of menadione or SpNO incubations (Fig. 3). Continued incubations of the cells in normal growth medium for another 20 h, resulted in increased GSH levels above the control levels (Fig. 3).

N.-E. Huseby et al. / Hotneostasis of Glutathione

ctr colls cells

ctr cells cells

men SpNO

men SpNO

Fig. 1. Cell viability after incubation of CC531 ceils with menadkme and SpNO. Cefls were incubated with 50 uM menadione for 15 min, or 200 pM SpNO for 4 h, and then in normal growth medium for a total time of 48 h. A: The number of viable cells after cultivation for 48 h in normal medium (ctr), or after exposure to menadione (men) or sperminNONOate (SpNO). B: The activity of LDH determined in cells floating in the medium after being harvested by centrifugation and lysed in Triton X-100. Data shown are means (+SD) of 4-8 measurements. * Indicates a significant difference to control cells.

B

100-

1 60-

T

control cells

cells + men +NAC

cells* SpNO +NAC

Figure 2. Induction of GGT after incubation of CC531 cells with menadione or SpNO. GGT activity was measured 48 h after incubations of the ceHs with menadione (B) or SpNO (C) in the medium as described in legend to Fig. 1. Incubations with 10 mM of the antioxktent N-acetylcysteine (NAC), were performed together with the menadione and SpNO exposures. Data shown are means (+ SD) of 4-6 experiments. * Indicates a significant difference in activity (p<0.05) compared to that of control cells (A).

175

176

N.-E. Huseby et al. / Homeostasis of Glutalhione

B

60

40

**

20-

1

i 4h 24h

4h 24h

control cells

cefls •>• men

24h eels + SpNO

Figure 3. Intracellular GSH levels after NO-exposure. Cete were incubated with menadk>ne or SpNO is indicated in Fig. 1, foJowed by 24 h in normal growth medkim. The amount of GSH was determined in eel rysates obtained 4 h after these exposures, or 20 h after. Control oate (ctr) were incubated in normal growth medum only. Data are from 3-5 experiments, and shown as mean + SO. * Indicates a significant difference (p«3.05) to GSH levels in control eels.

GSH depletion alone does not induce GGT To test whether GGT will accompany GSH depletion, the cells were treated with diethyhnaleate (500 uM diethylmaleate for 3 h) and incubated for 24 h in cystine-depleted medium. Both procedures resulted in a significant reduction in the GSH level (Fig. 4A), however, they were not followed by increased GGT activity (Fig. 4B). This indicates that there is no clear correlation between depletion of GSH, and induction of GGT. GGT prevents GSH depletion and cell death in cystine depleted medium To confirm that GGT may supply the colon carcinoma cells with cysteine was studied in medium without cystine but containing GSH. Incubating cells in cystine-depleted medium resulted in a strong reduction in intracellular GSH (Fig. 4A and 5), but this was prevented when GSH was added to the medium (Fig. 5). Inhibition of GGT activity with acivicin, which reduced the activity by more than 90%, or inhibiting the cyst(e)ine transporters by homocysteic acid and alanine, resulted in GSH reduction in spite of the added GSH (Fig. 5). Incubation of cells in cystine-depleted medium also resulted in a significant increase in apoptotic cells as quantified by the Cellular DNA Fragmentation ELISA. This number was significantly reduced when GSH was added to the medium (Fig. 6). Cells preincubated with SpNO show an increased uptake rate of cysteine To test whether cells with induced GGT obtained an increased availability of cysteine, we measured the uptake of cysteine. This was performed in cystine-depleted medium, to which was added 3SS-radiolabelled GSH. The uptake was almost linear during 8 h and was blocked by inhibiting GGT with acivicin (Fig. 7). The cysteine uptake was almost doubled in cells that had

N.-E. Huseby et al. / Homeostasis of Glutathione

177

been preincubated 4 days earlier with SpNO (Fig. 7), the GGT activity level of these cells was about twice the control level.

ctr

+OEM -cys

ctr

+DEM -cys

Figure 4. Cellular GSH level and GGT activity after cultivating cells with DEM and in cystinedepleted medium. Cells were incubated with 500 pM diethylmaleate (DEM) for 3 h, or in cystinedepleted medium for 24 h, and harvested for GSH measurements. GGT activity was measured after contiuned cultivation in normal growth medium for a total period of 48 h. Control cells (ctr) were incubated in normal growth medium only. Data are from 3-4 experiments, and shown as mean + SD. * Indicates a significant difference (p<0.05) to control cells.

-~.

60

i

*OSH +ala/HCA

Figure 5. Cellular GSH level after incubation in cystine-depleted medium. GSH was measured in the CC531 cells after 24 h incubation in control medium (ctr), in cystine-depleted medium (-cys), or in mis medium with 50 uM GSH (-cys/+GSH), with GSH and acMoin (-cys/+GSH+AC), or with GSH and alanin/homocysteic acid (-cys/+GSH/+ala/HCA). Data are from 3-6 experiments and shown as mean + SD. * Indicates a significant difference (p<0.05) to GSH levels in control cells.

178

N.-E. Huseby et al. / Homeostasis of Glutathione

Figure 6. Apoptosb in CC531 cells after incubation in cystine depleted medium. Cete were incubated in cystine-depteted medum (cys-) and in cystine-depleted medium enriched with 50 MM GSH (cys-/+GSH) for 24 h, and the amount of apoptptic cete was quantified using the DMA fragmentation ELJSA kit Control cete (ctr) were incubated in normal growth medum. Data shown are means (+SD) from 3-4 experiments. * Indicates a significant dfference (p<0.05) from cete in standard medium.

Cysteine uptake 200

150-

I I control eels mi GSwSt pfwncuoateo 100-

50

0J 0

4

8

Incubation period (h)

Figure 7. Cysteine uptake during incubation with radiolabeltod GSH. CC531 cete, either control cete or cete preincubated 4 days earlier with 200 fiM SpNO for 4 h fbfowed by cultivation in normal growth medum, or control cete with acivicin, were incubated with 36S-GSH and glycylgtycine in cystine-depleted medum. The cete were harvested in NaOH-SDS and the radoactivty was counted in a scintillation counter. The data are means of 3-4 measurements.

179

N.-E. Huseby el ai. /Homeostasis of Glutathione

Sublines 0/CC531 cells with higher malignancy have increased levels ofGSH but divergent GOT activity Two sublines of the CC531 cells is produced that can be characterized as more malignant than the standard cell line (termed LO). These are the L6 cells which have been cloned after repeated in vivo passages, and have a considerably higher metastatic growth rate in liver of syngeneic rats, and the RL4 cells which have a high resistance to cisplatin. A comparison of the GSH level and GOT activity in the three cell lines showed that GSH was considerably higher in both L6 and RL4 sublines compared to the control cells (Fig. 8A) whereas GGT activity in L6 cells was not different from the control cells, and that the level in the RL4 cells was only moderately higher (Fig. 8B). This does not support that GGT is consistently increased in more malignant cells, but that GSH may be so.

4. Discussion We have shown that the activity of GGT in colon carcinoma cells is induced after exposure to peroxides and nitroc oxide, and that the cells, when cultivated in cystine depleted medium, benefit from the enzyme in maintaining the intracellular level of GSH. Furthermore, cells with upregulated GGT are able to increase the degradation of extracellular GSH which results in a higher rate of cysteine uptake. The GGT activity will thus add to the protective measures of tumor cells during nitrosative and oxidative stress particularly at low extracellular concentrations of cysteine.

in 160

1

D I £.100

lui X

LO

US

RL4

CC531 subline

LO

L«

RL4

CC531 sublfne

Figure 8. Cellular GSH level and GGT activity in sublines of CC531 cells. Sublines of CC531 with higher ability to metastasise (L6 cells) or increased resistance to cisplatin (RL4 cells) were compared to the standard control cells from which they were derived (LO cells). The cells were harvested for GSH and GGT measurements. Date are from 3-4 experiments, and shown as mean + SD. * indicates a significant difference (p<0.05) to control cells.

180

N. -E. Huseby et al. / Homeostasis of Glutathione

There appears to be a rather long series of oxidative agents that upregulate GOT. We and others have earlier reported that GGT is induced by oxidative stress, such as exposure of hydrogen peroxide and quinones to cells, in a way that partly reflects the present results [1720,22]. The mechanism behind both the oxidative and the nitrosative induction of GGT remains to be evaluated, but the inducingagents may be free radicals that are produced during either stress procedure. In support for this are the effects of antioxidants which in part blocked the inductions and our findings that GSH depletion alone did not induce GGT [19,20]. During such situations, ROS may be generated, but apparently not sufficiently to induce GGT in the CC531 cells (data not shown). The genetic regulation of GGT is complex. In rats, the enzyme is coded for by one gene but 7 mRNA isotypes have been described. They have identical coding regions but differ in the 5'-untranslated regions. The presence of several promoters and regulatory elements strongly indicates that the enzyme is subjected to a strict regulation (for a recent review, see [23]). They may also explain how several mechanisms can upregulate GGT during oxidative and nitrosative stress, drug exposure or buryrate mediated differentiation [16-20,24]. GGT acts as a glutathionase under in vivo like conditions, that is at low concentration of cyst(e)ine [12,13,16]. The enzyme initiates the degradation of extracellular GSH and the level of GGT activity has been shown to correlate to the rate of cysteine uptake [12,13,15,16,25]. GSH is of significant importance in cellular defence against oxidative and nitrosative stress and different cell types use different pathways to maintain the intracellularGSH level [10,26]. Data indicate that the overall capability of cells to maintain a critical amount of GSH determines the susceptibility to NO-induced cell death [26]. Acute exposures of cells to NO frequently result in transient depletion of GSH, followed by GSH elevations [9,17,28], supporting the present data. This may be part of the adaptive response reported for endothelial cells after NO exposure; elevated GSH synthesis was found to coincide with the induction of Y-glutamylcysteine synthetase and increased uptake of cystine through the Xj" aminoacid transport system [28,29]. The colon carcinoma cell line CC531, established after chemical carcinogenesis in rat [21 ], has a relatively high GGT activity [19]. We found that GGT can supply the cells with cyst(e)ine by hydroly sing extracellular GSH when the cells were incubated in medium with low cystine, and thus confirmed data from transfected lymphocytes 13. GGT can in this way add to the protection of tumor cells against oxidative and nitrosative stress.

References 1. 2. 3. 4.

H. Sies. Glutathione and its role in cellular functions. Free Rad. Biol. Med. 27 (1999) 916-921. AG. Hall. The role of glutathione hi the regulation of apoptosis. Eur. J. CBn. Invest. 29 (1999) 238-245. JH. Doroshow, Glutathione peroxidase and oxidative stress. Toxicol. Lett. 82-83 (1995) 395-398. TE. Meyer.HQ. Liang, AR. Buckley, et al. Changes in glutathione redox cycling and oxidative stress response in the malignant progression of NB2 lymphoma cells. Int. J. Cancer 77 (1998) 55-63. 5. MJ. Anasagasti, JJ. Martin, L Mendoza et al. Glutathione protects metastatic melanoma cells against oxidative stress in the murine hepatic microvasculature. Hepatotogy 27 (1998)1249-56. 6. J. Carretero, E. Obrador, JM. Esteve, et al. Tumoricidal activity of endothelial cells. Inhibition of endothelial nitric oxide production abrogates tumor cytotoxicity induced by hepatic sinusoidal endothetium in response to B16 melanoma adhesion in vitro. J. Biol. Cnem. 276 (2001) 25775-82. 7. I. loannidis, H. de Groot. Cytotoxicity of nitric oxide in Fu5 rat hepatoma cells: evidence for co-operative action with hydrogen peroxide. Biochem. J. 296 (1993) 341-5.

N. -E. Huseby el al. I Homeostasis of Glutathione

181

8. HH. Wang, AR. Mcintosh, BB. Hasinoff, et al. B16 melanoma cell arrest in the mouse liver induces nitric oxide release and sinusoidal cytotoxicity: a natural hepatic defense against metastasis. Cancer Res. 60 (2000) 58629. 9. YS. Ho, HM. Lee, TC. Mou, YL. Wang, JK. Lin. Suppression of nitric oxide-induced apoptosis by N-acetyl-Lcysteine through modulation of glutathione, bcl-2, and bax protein levels. Mol. Carcinog. 19 (1997) 101-13. 10. V. Umansky, M. Rocha, R. Breitkreutz, et al. Glutathioneis a factor of resistance of Jurkat leukemiacells to nitric oxide-mediated apoptosis. J. Cell Biochem. 78 (2000) 578-87. 11. JB. Whitfield. Gamma glutamyl transferase. Crrt. Rev. Clin. Lab. Sci. 38 (2001) 263-355. 12.MH. Hanigan, WA. Ricketts. Extracellular glutathione is a source of cysteine for cells that express gammaglutamyl transpeptidase. Biochemistry 32 (1993) 6302-6. 13. DR. Karp, K. Shimooku, PE. Lipsky. Expression of gamma-glutamyl transpeptidase protects ramos B cells from oxidation-induced cell death. J. Biol. Chem. 276(2001)3798-804. 14. A. Komlosh, G. Volohonsky, N. Porat, et al. Gamma-glutamyl transpeptidase and glutathione biosynthesis in non-tumorigenic and tumorigenic rat liver oval cell lines. Carcinogenesis 22 (2001) 2009-16. 15. ME. Rajpert-DeMeyts, M. Shi, M. Chang, et al. Transfection with gamma-glutamyl transpeptidase enhances recovery from glutathione depletion using extracellular glutathione. Toxicol. Appl. Pharmacol. 114 (1992) 56-62. 16.Z. el-akawi, J. Zdanowicz, PJ. Creaven, hM Abu, R. Perez, L. Pendyala. Induction of gamma-glutamyl transpeptidase mRNAby platinum complexes in a human ovarian carcinoma cell line. Oncol.Res. 8 (1996) 41523. 17. MO. Ripple, PA. Pickhardt, G. Wilding. Alteration in gamma-glutamyl transpeptidase activity and messenger RN A of human prostate carcinoma cells by androgen. Cancer Res. 57(1997)2428-33. 18. RM. Liu, MM. Shi, C. Giulivi, HJ. Forman. Quinones increase gamma-glutamyl transpeptidase expression by multiple mechanisms in rat lung epithelial cells. Am.J.Physiol. 274 (1998) L330-L336 19.O. Borud, B. Mortensen, IM. Mikkelsen, P. Leroy, M.Wellman, NE. Huseby, Regulation of gammaglutamyltransf erase in cisplatin-resistantand -sensitivecolon carcinomacells after acute cisplatin and oxidative stress exposures. Int. J. Cancer 88 (2000) 464-8. 20. IM. Mikkelsen, B. Mortensen, Y. Laperche, NE. Huseby. The expression of y-glutamyttransferase in rat colon carcinoma cells is distinctly regulated during differentiation and oxidative stress Mol. Cell. Biochem. 2002 (in press). 21. RL Marquet, DL Westbroek, J. Jeekel. Interferon treatment of a transplantable rat colon adenocarcinoma: importance of tumor site. Int. J. Cancer 33 (1984) 689-92. 22. A. Kugelman, HA. Choy, R. Liu, MM. Shi, E. Gozal, HJ. Forman. gamma-Glutamyl transpeptidase is increased by oxidative stress in rat alveolar 12 epithelial cells. Am.J.Respir.Cell Mol.Biol. 11 (1994) 586-92. 23. N. Chikhi, N. Holic, G. Guellaen, Y. Laperche. Gamma-glutamyttranspeptidase gene organization and expression: A comparative analysis in rat, mouse, pig and human species. Comp. Biochem. Physio). B. 122 (1999)367-380. 24. H. Oguchi, F. Kikkawa, M. Kojima, et al. Glutathione related enzymes in cis-diamminedichloroplatinum (II)sensitive and-resistant human ovarian carcinoma cells. Anticancer Res. 14 (1994) 193-200. 25. R. Roozendaal, E. Vellenga, MA. de Jong, et al. Resistance of activated human Th2 cells to NO-induced apoptosis is mediated by gamma-glutamyltranspeptidase. Int. J. Immunol. 13 (2001) 519-28. 26. D. Berendji, V. Kolb-Bachofen, KL Meyer, KD. Kroncke. Influence of nitric oxide on the intracellular reduced glutathione pool: different cellular capacities and strategies to encounter nitric oxide-mediated stress. Free Rad. Biol. Med. 27 (1999) 773-780. 27. BJ. Buckley, AR. Whorton. Adaptive responses to peroxynitrite: increased glutathionelevels and cystine uptake in vascular cells. Am J Physiol Cell Physiol 279 (2000) C1168-C1176 28. H. Li, ZM. Marshall, AR. Whorton. Stimulation of cystine uptake by nitric oxide: regulation of endothelial cell glutathione levels. Am. J. Physiol. 276 (1999) C803-C811 29. D. Moellering, McAndrew AJ, RP. Patel, et al. The induction of GSH synthesis by nanomolar concentrations of NO in endothelial cells: a role for gamma-glutamylcysteine synthetase and gamma-glutamyl transpeptidase. FEBS Lett. 448 (1999) 292-6.

182

Thiol Metabolism and Redox Regulation of Cellular Functions A. Pompella et al. (Eds.) IOS Press, 2002

The Importance of gamma-Glutamyl Transferase in Lung Glutathione Homeostasis and Antioxidant Defense Martin JOYCE-BRADY, Yue LIU, Robert E. MARC and Jyh-Chang JEAN The Pulmonary Center, Boston University School of Medicine, USA

1. Introduction An extensive literature now supports a fundamental role for glutathione in the antioxidant protection of the lung and its gas exchange surface [1]. However, the physiological significance of lung glutathione turnover by the extracellular catabolic enzyme gammaglutamyl transferase (GGT) has remained controversial. The function of GGT in the lung has been a major research focus in my laboratory. Our studies initially focused on normal rat lung both at different stages of development [2-4] and in different conditions of oxidant stress [5]. More recently, these studies have been complemented by experiments with a newly defined model of GGT deficiency, the GOT*""1 mutant mouse [6,7]. The cumulative information we have generated from this work in rat and mouse, that correlates the cellular sites of GGT gene expression with that of GGT protein and enzyme activity, supports a critical biological role for GGT in lung glutathione homeostasis and epithelial cell antioxidant defense. Our initial studies in rat lung revealed key insights into epithelial cell biology, surfactant function, intercellular glutathione transport, the role of GGT in the lung response to oxidant stress, and the role of oxygen as a regulator of gene expression in the lung. The GGT*1"11 mutant mouse provided a model in which to test some of these observations and to more specifically define the role of GGT in the lung. In particular, we find that the absence of GGT expression in the mutant mouse lung severely limits glutathione availability and induces oxidant stress in normal oxygen conditions in both a subset of epithelial cells and in the surfactant that bathes cells at epithelial surfaces. Injury in these GGT-deficient lung epithelial cells is accelerated in the presence of hyperoxia with GOT*""1 mutants dying more rapidly compared to normal mice [8]. Our study of the GOT6""1 mouse also led us to identify novel truncated GGT protein isoforms that are generated by alternative splicing of GGT mRNA. While the intact GGT protein is expressed on the plasma membrane, these truncated isoforms are retained in the endoplasmic reticulum (ER). Surprisingly, both the native GGT and one of its ER isoforms can mediate an endoplasmic reticulum stress response demonstrating a new function for GGT [9]. Taken together, our results suggest that GGT expression is crucial for lung antioxidant defense and may serve a number of important homeostatic functions in lung epithelial cells. We believe that further investigation of lung GGT gene regulation and protein expression will allow us to identify new targets to design therapies for protecting both the developing and the mature lung and its epithelium against injury from oxidant stress.

M. Jovce-Brady et at. /The Importance of Gamma-Glutamyl Trans/erase

183

2. The GOT Gene and the Lung A summary of the GGT gene structure that is relevant to its expression in the lung is presented in Figure 1. The rat and the mouse GGT genes are very similar in organization and regulated by multiple alternative promoters [10]. The three most proximal promoters, PI, P2 and P3 are utilized under different conditions in the lung and result in the transcription of three different mRNA species. Each transcript has a unique 5' untranslated region, corresponding with alternative promoter usage, followed by a common 5' region and coding domain (ATG denotes the start codon). The protein is translated as a single chain propeptide that is enzymatically inactive [Rebecca Hughey, personal communication, and 11]. GGT is a type II integral membrane protein as the uncleaved amino-terminal signal anchor (SA) orients the translocation of GGT into the ER, leaving the amino terminus in the cytosol and the glycosylated ectodomain with the active site residues in the lumen [12]. It is now apparent that GGT is a likely member of the family of "N-terminal nucleophile hydrolases" [11], as proposed in previous studies [13]. The new amino terminal threonine produced by autocatalytic cleavage of the GGT into a large (L) and small (S) subunit heterodimer is also the catalytic nucleophile for the enzyme. The heterodimer is stably associated, and the catalytic site for hydrolysis of glutathione includes residues from both subunits.

Pi

\x< 1

—/ ATG

RNA

AAAAA

Protein Enzyme Glutathione + H2Q

Glu + CysGly

Fig. 1. Schematic of GGT gene. P1, P2, P3 denote alternative promoters and the corresponding mRNAs. SP marks signal peptide in primary translation product and vertical black lines are glycosylation. SA is retained membrane signal anchor with L as large and S as small subunits after processing. GGT initiates the metabolism of glutathione.

184

M. Joyce-Brady et al. / The Importance of Gamma-Glutamyl Trans/erase

GGT is highly expressed on the surface of cells with secretory or absorptive activity, such as epithelia. GOT activity is also present in epithelial cell secretions [14]. A hydrophilic form resulting from proteolytic removal of the transmembrane domain in hepatocytes accounts for the GOT normally found in the blood [15], while the amphipathic form found in bile duct is due to its solubilization by the detergent action of secreted bile salts [16]. Our own work has shown that GGT in lung surfactant retains its membrane anchor and bilayer association consistent with a secretory mechanism for GGT release from alveolar epithelial type 2 cells [2]. GGT activity within epithelial secretions is postulated to regulate the turnover and the size of both intracellular and extracellular glutathione pools. The results of studies in two recent models of GGT deficiency, the GGT11'1111 mouse [17] and the GOT""" mouse [6], have established the importance of this enzyme in mammalian glutathione homeostasis and cysteine supply. Our studies focused on the importance of GGT expression to glutathione homeostasis in the lung.

Surfactant

T2Cell

Fig. 2. Lung T2 cell GGT expression. Epithelial T2 and T1 cells and the alveolar macrophage (MO) are denoted and the results of Northern blot for GGT mRNA expression along with RT-PCR with P3 specific primers shown at left (Lv = liver. K = kidney, T2 = lung T2 cell, Lg = lung, H = heart). Protein immunoprecipitation is shown at right suing nonimmune (Mis) or immune (Is) serum. Enzymatically active GGT heterodimer is shown in secretory lamellar body of T2 cell and in association with extracellular lung surfactant.

M. Joyce-Brady et at. /The Importance of Gamma-Glutamyl Trans/erase

185

3. Alveolar Epithelial Type 2 Cell GGT and Surfactant Controversy over the importance of GGT and glutathione metabolism in the lung stemmed from a deficit of knowledge about lung GGT expression. The level of GGT activity in the lung as a whole was known to be very low compared to other organs, especially the kidney. So glutathione metabolism in the lung was considered physiologically insignificant [18]. Studies focused solely on the presence of lung GGT enzyme activity or lung GGT protein suggested several different cell types [19-21] as potential GGT sources, including the flat alveolar epithelial type 1 (Tl) cell [22]. Since the focus of my laboratory is lung alveolar epithelial cell biology, this observation suggested that GGT could serve as a new tool to investigate the process of differentiation as the cuboidal alveolar epithelial type 2 (T2) cell differentiates into the flat Tl cell. In order to study this process precisely, we integrated the expression of GGT mRNA [23] with that of protein and enzyme activity. Contrary to our working hypothesis, analysis of RNA from freshly isolated lung alveolar epithelial Tl cells, type 2 (T2) cells and alveolar macrophage cells (MD) revealed expression of GGT mRNA solely in the T2 epithelial cell (Figure 2). PCR analysis with upstream primers targeting the unique 5' UT GGT mRNA regions showed that this T2 cell GGT transcript was derived from promoter 3, a promoter initially characterized from the rat liver [24]. Metabolic labeling experiments confirmed that T2 cells synthesize GGT protein and enzyme assays showed an enrichment of GGT activity in isolated T2 cells over that in whole lung (Figure 3). This data provided firm evidence that GGT was expressed by the T2 cell.

GGT Enzyme Source Kidney Lung Heart

T2 cell Ti cell Macrophage

Specific Activity 1860 ±435 68±15

15.0 ± 3.4 1,4 ± 0.6 1.6 ± 0.5

Lung Lavage

Surfactant-enriched fraction SurfectaJit-depleted fraction

46.2 i 5.6 1.3 ± 0.2

Fig. 3. GGT enzyme activity. Values are specific activities in organs, lung cells and lung lavage fractionated into surfactant-enriched or depleted fractions.

However, we did detect a measurable level of GGT activity in the Tl cell and the alveolar macrophage, despite the absence of GGT mRNA expression. To explain this observation, we performed several additional experiments to show that the GGT activity in these cells could be derived from the T2 cell via surfactant, the lipoglycoprotein secretory

186

M. Joyce-Brady el al. / The Importance ofGamma-Glutamyl Transferase

product of the T2 cell. First, T2 cells could be stimulated to release GGT activity in parallel with surfactant lipid. Second, lung lavage fluid contains GGT activity that, upon centrifugation, partitions with the surfactant enriched fraction. The specific activity of GGT in this fraction is about 7-fold enriched over that of whole lung. Lastly, and in collaboration with Dr. Rebecca Hughey, GGT activity obtained from lung lavage fluid also partitions into the detergent phase of Triton X-l 14, like amphipathic GGT in renal brush border membranes [2]. GGT activity can be released from either the surfactant-enriched fraction or the detergent phase of Triton X-l 14 by the enzyme papain, which cleaves the GGT signal anchor and renders the protein hydrophilic [25]. Hence the GGT activity associated with lung surfactant results from the amphipathic protein. These results suggested to us that lung surfactant can serve a novel function by acting as a vehicle to redistribute this amphipathic molecule throughout the entire gas exchange surface of the lung. Since surfactant lines the entire surface of the lung, this also implied that surfactant-associated GGT protein could have a broad impact on the biology of other lung cells, in addition to the T2 cell. For instance, glutathione is abundant in the lung lining fluid [18]. The presence of GGT activity suggests that glutathione turnover occurs in this pool. In fact, we showed that when the lung is exposed to an inhaled oxidant, there is a dramatic accumulation of surfactant-associated GGT activity [5]. Therefore, the response of the lung to this inhaled oxidant stress may involve an acceleration in the turnover of glutathione within this extracellular pool.

4. GGT Ontogeny in Lung Epithelial Cells The ontogeny of GGT in the lung is a tale of two different epithelial cell types, the alveolar epithelial type 2 (T2) cell and the non-ciliated bronchiolar epithelial (Clara) cell [3]. GGT activity accumulates at two distinct periods during lung development (Figure 4). The first occurs in the fetal lung, late in gestation, co-incident with the accumulation of surfactant phospholipids and several antioxidant enzyme activities. The level of activity, however, is only a fraction of that seen in the adult lung. This first peak results from activation of GGT gene expression in the alveolar epithelial type 2 cell (T2). In fact, GGT ontogeny from the late fetal through the early postnatal periods is a feature solely of the T2 cell. This suggests that glutathione metabolism may be important in the alveolar region of the fetal lung and at the gas exchange surface of the perinatal lung. When we used PCR to examine the pattern of GGT promoter usage in the fetal lung [4], we found that GGT promoters 1 and 2 were active, in addition to GGT promoter 3. A change in this pattern of GGT promoter utilization became apparent almost immediately after birth with the disappearance of the GGT P2 mRNA within the first 24 hours of breathing oxygen. T2 cells, isolated from adult lung and returned to the level of hypoxia of the fetal lung, re-express GGT P2. In contrast, GGT promoter 1 remained active in the newborn lung, but was inactivated by day 10 after birth. GGT PI can be re-activated if the adult lung is exposed to an inhaled oxidant [5]. The persistent usage of GGT promoter 3, under normal oxygen conditions, appeared to be due to its activation by oxygen. This differential regulation of the alternative GGT promoters in the lung suggested a mechanism to maintain lung GGT gene expression, and glutathione metabolism, over a wide range of oxygen concentrations. In addition, the change in the oxygen environment at birth may function to regulate the pattern of gene expression in the postnatal lung. Genes that are essential for postnatal development will be activated, and genes important for fetal lung development will be repressed. These data suggest

M. Joyce-Brady et al. /The Importance of Gamma-Glutamyl Trans/erase

187

that birth is a natural experiment with hyperoxia and that environmental oxygen could impact the course of postnatal lung development. However, the biological role of this regulation of the alternative GGT promoters remains obscure since the identical protein product is generated from all three promoters. Alternatively, promoter usage may be linked to mRNA function if sequences within the unique 5' untranslated regions of the GGT mRNA species can be shown to differentially regulate mRNA translation or stability in lung cells under different oxygen conditions.

?. ' f;

. *

6

A C T I V I T Y

Bst

5 4

3

*.

$

S)

• Clara cell

T2cell

•'

+ T2 cell

2

P3 P2

* 0

PI , - v " . • ' N pi + pa

*

16 18 20 22

FETAL days

f

. . "

4 1

2

4

6

8

2 4 6

8

POSTNATAL* Inhaled Oxidant wwk? | rt^ys

Fig. 4. GOT Ontogeny in Lung Epithelial Cells. Ordinate shows GGT activity (nmol/min/mg protein) and abscissa time during development GGT ontogeny in specific lung epithelial cells is shown in grey. Activity of alternative promoters P1, P2, P3 is listed in black.

The second peak of GGT activity occurs abruptly during the late postnatal period and raises the level of GGT activity to that seen in the adult lung. This peak results from activation of the GGT gene in an epithelial cell of the distal airways in the lung. This cell is the nonciliated bronchiolar Clara cell and it becomes the primary cellular site of expression of GGT mRNA, protein (Figure 5) and activity in the lung (3). This rise in GGT activity appears to correlate with an accumulation of GGT heterodimer in the lung, in contrast to the alveolar epithelial T2 cell where newly synthesized propeptide appears to accumulate in excess of heterodimer. It is possible that GGT protein serves a different function in these the two different epithelial cell types, perhaps related to the activation of GGT enzyme activity. Certainly, the bronchiolar Clara cell has a high demand for glutathione as a substrate for xenobiotic metabolism, and for antioxidant defense since it is located in small airways that are

188

M. Joyce-Brady et al. / The Importance of Gamma-Glutamyl Transferase

the primary sites for deposition of inhaled particulate matter. Expression of enzymatically active GGT heterodimer on its cell surface could ensure access to the pool of glutathione in the lung epithelial lining fluid. In support of this, GGT gene expression is highly induced in this epithelial cell following exposure of the lung to an inhaled oxidant gas, thereby increasing cellular access to extracellular glutathione. Taken together, these observations suggest that GGT expression is important for the bronchiolar Clara cell in the mature lung and imply two things: first, loss of lung GGT expression should make this cell particularly vulnerable to glutathione depletion and injury by oxidants; and second, since activation of Clara cell GGT expression is delayed until the late postnatal period, the bronchioles of the perinatal lung may be particularly susceptible to oxidants and inhaled particles.

Fetal Lung

Postnatal Lung

T2

Clara

Fig. 5. Immunolocalization of GOT protein in lung parenchyma during development. GGT protein product (brown color) localizes to alveolar T2 cell in rat lung at fetal day 21 and postnatal day 14 (black arrows), but only bronchiolar cells in the postnatal lung (red arrows).

5. Consequences of Lung GGT Deficiency in the GGT*"1 Mouse The GOT6™1 mutant mouse provided a model in which to test some of these observations and to determine if GGT expression was important in the hing. The GOT*™1 mouse is genetic model of GGT deficiency that was generated by randomly inducing point mutations in the mouse genome with ethylnitrosourea (enu) then selecting progeny for aminoaciduria In the case of the GGT60"1 mouse, the aminoaciduria was glutathionuria. This resulted from the nearly complete inactivation of GGT enzyme activity in the kidney. A systemic inactivation of GGT

M. Joyce-Brady et al. /The Importance of Gamma-Glutamyl Trans/erase

189

gene expression was suggested by the presence of growth retardation, premature mortality, infertility and cataracts [6]. The GOT6011' phenotype was very similar to that of the GGTml/ml mouse, a model of complete GGT gene inactivation produced by targeted mutagenesis of the mouse GGT gene [17]. The glutathionuria in the GGTml/ml mouse was of a much greater magnitude than that in the GGT61"11 mouse and caused a deficiency in plasma cysteine content, systemic cysteine supply and glutathione content in many tissues. In contrast, a small residua of GGT activity persisted in the kidney of the GGT60"1 mouse and the level of glutathionuria was less severe. Plasma cysteine content remained normal though a decrease in urinary excretion of taurine, a cysteine metabolite, suggested a relative deficiency in the supply of Cysteine or other sulfur-containing amino acids. Although the phenotype is similar in these two mouse models, the manifestations of GGT deficiency in the GOT6™1 mouse are less severe than that in the GGTml/ml mouse. In order to study the GGT6""1 mouse, we first had to characterize the site of the point mutation in order to develop a genotyping strategy to breed the mice in my laboratory. Since GGT is a single copy gene in the haploid mouse genome, and the phenotype suggested widespread GGT deficiency, we suspected that the point mutation was in the coding domain. We located the mutation in the first coding exon, where an AT-TA transversion caused a leucine codon (TTG) to be replaced by a stop codon (TAG). Introduction of this premature translation termination codon caused a secondary decrease in the steady state level of GGT mRNA in the kidney, a phenomenon known as a low RNA phenotype. In addition, premature termination of GGT protein synthesis produced a severely truncated oligopeptide lacking all of the amino acid residues required for enzyme activity. Together, these contributed to the loss of nearly all GGT activity. A tiny residua of GGT activity persisted in the kidney, which normally expresses GGT at very high levels. We proposed that this could result if some GGT protein synthesis was initiated at the downstream methionine codon, 117, which is also preceded by a Kozak consensus sequence. Despite the residual GGT activity, the kidney in the OQjenui mouse dispiays evidence of oxidant stress as heme oxygenase-1 mRNA is induced by 4-fold, and Cu, Zn SOD mRNA by 3-fold compared to that in normal kidney [7]. We then examined the GGTmul mouse lung for evidence of oxidant stress. To determine this, we assessed: 1) the pattern of lung GGT promoter usage, 2) the status of lung glutathione content and redox ratio, 3) the presence of a 3-nitrotyrosine signal in lung cells by immunohistochemistry, and 4) the response of these mice to an environment of >95% oxygen [8]We used RT-PCR with GGT mRNA subtype specific primers to assess activity of the three GGT promoters. GGT promoter 3 was active in the normal mouse kidney and the lungs of normal and GGT61"11 mice and normal rat. GGT promoter P2 was only active in the normal mouse kidney. But GGT promoter 1 was active in normal mouse kidney and in GOT6""1 lung. Since GGT PI promoter is activated by oxidant stress in normal lung, these results are consistent with oxidant stress in the GGT6"1*1 lung. Total lung glutathione content, measured by the recycling method of Teitze, was only minimally decreased in homogenates of GGTenuI compared to normal lung (Figure 6). This likely reflected the adequacy of lung cysteine supply in the GGT6""1 mouse. However, the fraction of GGT6™1 lung GSSG (oxidized glutathione), assessed by HPLC in collaboration with Dr. Lou Ann Brown, was increased more than 3-fold. This suggested oxidant stress in sub-populations of lung cells, cells where glutathione homeostasis was likely dependent on lung GGT expression. Therefore we assessed glutathione content in specific lung cells and lung lining fluid.

190

M. Joyce-Brady et al. /The Importance ofGamma-Glutamyl Transferase

GOT60111

Normal

0.76 ±0.15

0.80 ±0.12

%GSSG

31±2

9±1

M<j> (nmol/tecelk)

%GSSG

0.88 ± 0.55 70 ± 6

5.1 ±1.0 10 ±2

ELFfrM) GSH/GSSG

1639 ± 164 3.3 ±0.26

834 ±190 6.9 ± 0.2

240 ± 52

61 ±2

Lung (nmoJ/xng protein)

Plasma (>iM)

Fig. 6. Glutathione Content Glutathione content and redox ratio were determined in the lung, the macrophage, the epithelial lining fluid (ELF) and the plasma. The ELF values were normalized using the urea dilution method.

Glutathione content can be measured readily in the alveolar macrophage and the lung lining fluid using simple lung lavage. Total glutathione content in the GGT*""1 macrophage, determined by HPLC, was decreased by almost 6-fold compared to the normal macrophage and the GGT^1 macrophage GSSG fraction was increased by 7-fold indicating intense oxidant stress. Macrophage glutathione content and the redox state of the cell were severely compromised by the absence of glutathione turnover in the alveolar lining fluid. Interestingly, the concentration of glutathione in the GOT6™11 alveolar lining fluid was actually increased 2-fold over that in normal lung. But it was also more oxidized as the GSSG content increased by 4-fold. These data are consistent with a role for surfactant-associated GGT in the turnover of the extracellular glutathione pool in the lung. This result mimics the glutathionemia seen in the GOT61"*1 plasma. Three questions arose from these studies. How did GSSG rise disproportionately over GSH? Is there a decrease in GSSG clearance over GSH in the GGT60"1 lung?. Would epithelial cells in the GOT61™1 lung be protected against oxidants by this expanded glutathione pool? Glutathione content in the mouse lung epithelium was not so simple to assess, particularly the non-ciliated bronchiolar epithelial cell population. A cell isolation procedure could only produce a partially pure population of these cells and might also perturb cellular glutathione content. Therefore, we assessed epithelial cell glutathione content in situ using an immunohistochemical technique in collaboration with Dr. Robert Marc at the Moran Eye Center of the University of Utah. Dr. Marc is an expert on the development and application of immunologic reagents to localize glutathione and its related amino acids in cells and tissues [26]. To validate this technique in the GGT6™1 mouse we first studied the liver. Biochemical measure of total glutathione content in a homogenate of the GOT*0"1 liver is decreased by 3-fold compared to that of normal liver [6]. The immunohistochemical method showed a decrease in glutathione signal intensity in the GGT cnul liver compared to normals and the calibrated glutathione signal intensity in the GOT*™11 hepatocyte was decreased by 4-

191

M. Joyce-Brady elal /The Importance of Gamma-Glutamyl Trans/erase

fold compared to the normal hepatocyte (175 uM vs 830 uM). Hence glutathione content correlated very well when assessed by these two different methods (Figure 7).

N O R M A L

G G T

e n ii 1

Fig. 7. Immunocytochemical localization of glutathione and 3-nitrotyrosine (3NT). The liver and the lung from normal (left) and GGT*nu1 (right) mice was examined for glutathione content in the top three panels and 3-NT in the bottom panel. For glutathione, black signal denotes abundance and grey signal depletion of cellular glutathione content. A normal and GOT8™1 lung bronchiole (Br) is compared at low and high power in the second and third panels. Dense black signal localizes to ciliated epithelial cell in both. Glutathione signal is less dense in normal Clara cell but grey in GOT"™1 Clara cell indicating glutathione depletion. The 3 NT signal (brown color) at the bottom shows sparse and weak signal in normal bronchiolar cells but dense and uniform signal in GGT8""1 bronchiolar cells.

This immunohistochemical technique was then applied to assess glutathione content in the lung, particularly epithelial cells in the bronchioles. Non-ciliated Clara cells are found in the bronchiole as well as a second epithelial cell type, the ciliated epithelial cell. In the normal lung, a glutathione signal was present in both types of epithelial cell, but it was more intense in the ciliated cell. The reason for this difference is not yet clear but it suggests a difference in glutathione utilization. In the GGT superenul lung, an intense glutathione signal was still present in the ciliated cell but the signal was very weak in the non-ciliated (Clara) cell. Hence, the GGT deficient Clara cell is glutathione depleted. Alveolar epithelial type 2 cells express much lower levels of GGT than Clara cells and a difference in glutathione content between normal and GGT6""1 T2 cells was not evident by this technique.

192

M. Joyce-Brady el al. /The Importance of Gamma-Glutamyl Transferase

To determine if glutathione deficiency in the GGT""1 Clara cell was associated with oxidant stress, we probed for the presence of 3-nitrotyrosine. These modified tyrosine residues result from the interaction of tyrosine with peroxynitrite and are stable products [27]. A 3nitrotyrosine signal (brown color in lower panel of Figure 7) was actually evident in some bronchiolar cells of normal lung but it was weak and sparse. This probably reflects the dynamic nature of cellular glutathione pools in bronchiolar cells. In contrast, the 3nitrotyrosine signal was more intense and uniform in the GOT™"1 bronchiolar epithelium, consistent with oxidant stress. Specificity of the signal was assured by co-incubating the primary antibody with 10 mM nitrotyrosine which eliminated the nitrotyrosine signal. As a 3nitrotyrosine signal was also evident in the alveolar macrophage, there was good correlation between loss of lung cell GGT expression with sites of decreased glutathione availability and oxidant stress.

N O R M A L

G G T e n

Fig. 8. Immunotocalization of glutathione (GTH) and heme oxygenace-1 (HO1). The lungs of normal and GGTerenul mice were examined after a 96 hour exposure to hyperoxia. Dense black signal in top panel indicates abundant glutathione (GTH) in cells and lining fluid of normal lung but grey signal indicates depletion of glutathione from cells and lining fluid in GGT*""1 lung. Heme oxygenase-1 expression (brown signal) localizes only to vascular cells in normal lung. But to vascular cells and the entire epithelial surface in the GGTn anu1 lung where severe injury and cell damage is evident in the bronchiolar epithelium. Even in normoxia (bottom panel), HO-1 was localized to vascular cells but only in the GGT"0"1 lung.

u 1

M. Joyce-Brady et ai / The Importance ofGamma-Glutamyl Trans/erase

193

To determine the consequences of change in epithelial cell redox state, we studied the response of the GOT6™1 mice to an environment of >95% oxygen and found that the GOT61"11 mice died more rapidly than the normal mice. Cellular injury and pulmonary edema developed more quickly in the GOT6™1 lung. This was quite evident when we compared normal and GOT611111 lung after 96 hours in hyperoxia (in Figure 8). GOT60"1 lung glutathione content was widely depleted by glutathione immunohistochemistry compared to the normal lung. In addition, heme oxygenase-1 protein expression was also widely induced throughout the entire epithelial surface of GGT6™1 lung but not in the normal lung. Interestingly, heme oxygenase-1 was already localized to vascular endothelial cells even in normal oxygen conditions but only in the GGT6""1 lung. So these cells were apparently under oxidant stress even prior to exposure to hyperoxia, and thus one would expect injury and pulmonary edema to develop more rapidly in the GOT6*1"1 iung ft is not yet clear whether this sensitivity resulted from loss of GGT at the endothelial cell surface or the circulating pool of GGT in the plasma. Taken together, our data supports an important role for GGT expression in the adult lung for antioxidant defense. Our present goals are three-fold: 1) to determine the consequences of GGT deficiency on lung epithelial cell and macrophage gene expression and function; 2) determine if GGT expression is important in the lung at birth under different oxygen conditions; and 3) begin to explore the mechanism by which GGT protects epithelial cells from oxidant stress. Of considerable interest is the question of whether GGT senses glutathione content or the cellular redox state, and whether it performs a function that is specific to epithelial cells. 6. New Role for GGT in the Endoplasmic Reticulum During our characterization of the point mutation that inactivated GGT gene expression in the GGT6""1 mouse, we identified four alternative splicing events in mouse GGT cDNA. Dr. Rebecca Hughey will discuss our detailed characterization of these events at this conference. I wish to use one of these events to introduce the concept that GGT may serve a new function in the epithelial cell. The first alternative splicing event that we found involved the insertion of 22 new bases into the mouse GGT cDNA. The insertion caused a frame shift and introduced a premature stop codon into the GGT cDNA so that a new GGT protein was encoded. This protein was truncated within the large subunit and included 14 novel amino acid residues at the C-terminus. The loss if all residues from the small subunit predicted an absence of GGT activity. Our initial impression was that ethylnitrosourea treatment introduced a point mutation in the GGTenul mouse that disrupted a constitutive splice site. However, a search of GenBank revealed that the same 22 base insertion was already described as a naturally occurring alternative splicing event in a human GGT cDNA [28]. Therefore, a mutation did not cause this event. Rather, cloning and partial sequencing of mouse GGT intron 7 revealed that the identical 22 base insertion sequence, preceded by a CAG, was shared between mouse and human GGT. This was an alternative splicing event that occurred within intron 7, so we called it GGTA7 (Figure 9). The corresponding human GGT protein product was never characterized. Therefore, Dr. Hughey and I embarked on a collaborative effort to characterized this GGT protein isoform. The GGTA 7 protein was a rather stable glycoprotein monomer that lacked GGT enzyme activity and was retained in the endoplasmic reticulum. This alternative splicing event was regulated in a developmental and tissue-specific fashion, suggesting that it served a physiologic function. When expressed in CHO cells, this isoform, as well as native GGT,

194

M. Joyce-Brady el al. /The Importance ofGamma-Glutamyl Trans/erase

mediated an endoplasmic reticulum stress response after the cells were challenged with a medium containing solely cystine, instead of a mixture of cysteine and cystine, instead of solely cysteine. Although we do not yet know the metabolic basis for this effect, the result suggests that GGT functions within the ER in a previously unknown fashion. This implies that GGT may actually serve a number of important homeostatic functions in the epithelial cell [9].

Protein mGGT GGTA7

Fig. 9. Schematic of Alternative Splicing Event in QGTA7. Cartoon shows site of alternative processing event at GGT gene in exon 7. The 22 base insertion in GGT mRNA is shown in grey italics (boxed) for mouse (Mm) and single nudeotide difference in human GGT mRNA (Hs) is listed below. Normal mouse GGT protein and truncated GGTd? protein isofbrm with its new C-terminus are compared at the bottom.

7. Questions Raised by our Research on Lung GGT Expression The goal of this workshop is to gain new insight into "Thiol Metabolism and Redox Regulation of Cellular Function." Therefore, the following is a summary of several questions about GGT gene expression that our work has raised and that should lead to new insights into the biologic functions of GGT. What is the biological role of alternative promoter usage in regulating lung GGT gene expression in the lung? Does it affect the stability or the translational efficiency of different GGT mRNA species in lung epithelial cells? Does alternative GGT promoter usage impact the pattern of alternative GGT mRNA splicing within the coding region? Or is this alternative splicing pattern regulated by the cellular redox state or the cellular content of glutathione of

M. Joyce-Brady et al. /The Importance of Gamma-Glutamyl Trans/erase

195

lung epithelial cells? Is there regulation of the autocatalytic activity that converts enzymatically inactive GGT propeptide into the active heterodimer? Such a mechanism could impact the rate of glutathione turnover which may prove important in the endoplasmic reticulum, at the cell surface or within epithelial cell secretions. In the lung, this could play a role in regulating glutathione turnover and the size of the glutathione pool in the lung lining fluid which protects the gas exchange surface against oxidants. It may also provide insight into different functions that GGT might serve in the Clara cell and the T2 cell. Lastly, how does GGT protect lung epithelial cells from oxidant stress? Does it sense glutathione directly or, alternatively, the cellular redox state? How important is this function in alveolar epithelial type 2 cells in the perinatal lung before and after birth?

Acknowledgments The authors acknowledges the dedicated efforts of the following individuals whose participation over time was integral to the completion of these studies. Dr. Rebecca P. Hughey at the Laboratory of Epithelial Cell Biology in the Department of Medicine, Renal/Electrolyte Division of the University of Pittsburgh School of Medicine, Dr. Lou Ann Brown in the Department of Pediatrics at Emory University School of Medicine, Dr. Mary Williams at The Pulmonary Center of Boston University School of Medicine, and Dr. Yuji Takahashi, now at the Tokyo University School of Life Sciences. The authors thank Rebecca P. Hughey and Jerome S. Brody for critical reading of the manuscript.

References 1. 2.

3. 4.

5.

6.

7.

8.

9.

Rahman, I., W. MacNee. Lung glutathione and oxidative stress: implications in cigarette smoke-induced airway disease. American Journal of Physiology. 277 (1999) L1067-L1088. Joyce-Brady, M., Y. Takahashi, S. M. Oakes, A. K. Rishi, R. A. Levine, C. L. Kinlough, R. P. Hughey. Synthesis and Release of Amphipathic _-glutamyl Transferase by the Pulmonary Alveolar Type 2 Cell. Journal of Biological Chemistry. 269 (1994) 14219-14226. Oakes, S. M., Y. Takahashi, M. C. Williams, M. Joyce-Brady. Ontogeny of Gamma-glutamyltransferase in the rat lung. American Journal of Physiology. 272 (1997) L739-L744. Joyce-Brady, M., S. M. Oakes, D. Wuthrteh, Y. Laperche. Three Alternative Promoters of the Rat GammaGlutamyl Transferase Gene Are Active in Developing Lung and Are Differentially Regulated by Oxygen after Birth. Journal of Clinical Investigation. 97 (1996) 1774-1779. Takahashi, Y., S. M. Oakes, M. C. Williams, S. Takahashi, T. Miura, M. Joyce-Brady. Nitrogen Dioxide Exposure Activates Gamma-Glutamyl Transferase Gene Expression in Rat Lung. Toxicology and Applied Pharmacology. 143 (1997) 388-396. Harding, C. O., P. Williams, E. Wagner, D. S. Chang, K. Wild, R. E. Colwell, J. A. Wolff. Mice with Genetic Gamma-Glutamyl Transpeptidase Deficiency Exhibit Glutathionuria, Severe Growth Failure, Reduced Life Spans, and Infertility. Journal of Biological Chemistry. 272 (1997) 12560-12567. Jean, J. C., C. O. Harding, S. M. Oakes, Q. Yu, P. K. Held, M. Joyce-Brady. Gamma-Glutamyl transferase (GGT) deficiency in the GGT8™1 mouse results from a single point mutation that leads to a stop codon in the first coding exon of GGT mRNA. Mutagenesis. 14 (1999) 31-36. Jean, J. C., Y. Liu, L. A. Brown, R. E. Marc, E. Klings, M. Joyce-Brady. The GGT deficient Mouse Lung Senses Oxidant Stress in Normoxia. American Journal of Physiology (2002): in Press at under 'articles in Press' Joyce-Brady, M., J. C. Jean, R. P. Hughey. Gamma-Glutamyltransferase and Its Isoform Mediate an

196

M. Joyce-Brady et al. / The Importance of Gamma-Glutamyl Transferase

Endoplasmic Reticulum Stress Response. Journal of Biological Chemistry. 276 (2001) 9468-9477. 10. Chikhi, N., N. Hoik:, G. Guellaen, Y. Laperche. Gamma-glutamyl transpeptkJase gene organization and expression: a comparative analysis in rat, mouse, pig and human species. Comparative Biochemistry and Physiology. 122 (1999) 367-380. 11. Suzuki, H., H. Kumagai. Gamma-glutamyltranspeptidase: a new member of Ntn-hydrolase superfamBy Tanpakushitsu Kakusan Koso. 46 (2001) 1842-1848. 12. Taniguchi, N., Y. Ikeda. Gamma-Glutamyl Transpeptidase: Catalytic Mechanism and Gene Expression. Advances in Enzymology and Related Areas of Molecular Biology. 72 (1998) 239-278. 13. Brannkjan, J. A., G. Dodson, H. J. Duggteby, P. C. Moody, J. L. Smith, D. R. Tomchick, A. G. Murzin. A protein catalytic framework with an N-terminal nucleophile is capable of self-activation. Nature. 378 (1995) 416-419. 14. Meister, A., S. Tate, O. Griffith. Gamma-Glutamyl Transpeptidase. Meth. Enzym. 77 (1981) 237-253. 15. Matsuda, Y., A. Tsuji, N. Katunuma. Membrane Bound Gamma-Glutamyltranspeptidase: Its Structure, Biosynthesis and Degradation. Advances in Enzyme Regulation. 21 (1983) 103-119. 16. Hirata, E., M. Inoue, Y. Morino. Mechanisms of Biliary Secretion of Membranous Enzymes: Bile Acids Are Important Factors for Biliary Occurrence of Gamma-glutamyltransferase and Other Hydrolases. Journal of Biochemistry. 96 (1984) 289-297. 17. LJeberman, M. W., A. L Wiseman, Z. Shi, B. Z. Carter, R. Barrios, C. Ou, P. Chevez-Barrios, Y. Wang, G. M. Habib, J. C. Goodman, S. L. Huang, R. M. Lebovitz, M. M. Matzuk. Growth retardation and cysteine deficiency in gamma-glutamyl transpeptidase-deficient mice. Proceedings of the National Academy of Sciences, USA. 93 (1996)7923-7926. 18. Cantin, A. M., S. L. North, R. C. Hubbard, R. G. Crystal. Normal alveolar epithelial lining fluid contains high levels of glutathione. Journal of Applied Physiology. 63 (1987) 152-157. 19. Rutenburg, A. M., H. Kim, J. W. Fischbein, J. S. Hanker, H. L. Wasserkrug, A. M. Seligman. Histochemical and Ultrastructural Demonstration of gamma-Glutamyl Transpeptidase Activity. Journal of Histochemistry and Cytochemistry. 17 (1969) 517-526. 20. Dinsdale, D., J. A. Green, M. M. Manson, M. J. Lee. The Ultrastructural immunotocalization of gammaglutamyltranspeptidse in rat lung: correlation with the histochemical demonstration of enzyme activity. Histochemical Journal. 24 (1992) 144-152. 21.Forman, H. J., D. C. Skelton. Protection of alveolar macrophages from hyperoxia by gamma-glutamyl transpeptidase. American Journal of Physiology. 259 (1990) L102-L107. 22. Ingbar, D. H., K. Hepler, R. Dowin, E. Jacobsen, J. M. Dunitz, L Nici, J. D. Jamieson. Gamma-Glutamyl transpeptidase is a polarized alveolar epithelial membrane protein. American Journal of Physiology. 269 (1995) L261-L271. 23. Cotoma, J., H. C. Pilot. Characterization and sequence of a cDNA done of gamma-glutamyKranspeptidase. Nucleic Acids Research. 14 (1986) 1393-1403. 24. Griffiths, S. A., M. M. Manson. Rat liver gamma glutamyl transpeptidase mRNA differs in the 5' untranslated sequence from the corresponding kidney mRNA. Cancer Letters. 46 (1989) 69-74. 25. Hughey, R.

P.,

N.

P. Curthoys.

Comparison of the

Size

and Physical

Properties of gamma-

GlutamyttranspeptkJase Purified from Rat Kidney Following Solubilizatkm with Papain or with Triton X-100. Journal of Biological Chemistry. 251 (1976) 7863-7870. 26. Marc, R. E. Mapping glutamatergic drive in the vertebrate retina with a channel permeant organic ion. Journal of Comparative Neurology. 407 (1999) 47-64. 27. Ischiropoutos, H. Biological tyrosine nitration: A pathophysiological function of nitric oxide and reactive oxygen species. Archives of Biochemistry & Biophysics. 356 (1998) 1-11. 28. Pawiak, A., E. H. Cohen, J. Octave, R. Schweickhardt, S. Wu, F. Bulte, N. Chikhi, J. Baik, S. Siegrist. G. Guellaen. An Alternatively Processed mRNA Specific for gamma-Glutamyl Transpeptidase in Human Tissues. Journal of Biological Chemistry. 265 (1990) 3256-3262.

Thiol Metabolism and Redox Regulation of Cellular Functions ,4. Pompella et al. (Eds.) 1OS Press, 2002

' 97

The Role of gamma-Glutamyltranspeptidase in the Metabolism and Cytotoxicity of 4-Hydroxynonenal-Glutathione Conjugate: Evidence and Hypothesis Milica ENOIUt§, Regine HERBER§, Pierre LEROY§, and Maria WELLMAN5 *Faculty of Pharmacy, University Carol Davila, 6 Traian Vuia, 70139 Bucharest, Romania ^ "Thiols etfonctions cellulaires", Univ. Henri Poincare Nancy 1, 54000 Nancy, France

1. Introduction Gamma-glutamyltranspeptidase [GOT (5-glutamyl)-peptide:aminoacid 5-glutamyl transferase, E.G. 2.3.2.2] is a well known enzyme which cleaves the v-glutamyl moiety of glutathione (GSH) and of GSH conjugates in general. GGT initiates the breakdown of GSH and provides therefore the amino acids precursors for the intracellular synthesis of GSH. In the case of GSH conjugates, the reaction catalysed by GGT is followed by removal of glycine by dipeptidase and N-acetylation by cysteine conjugate-N-acetyltransferase, thus forming mercapturic acids as end products. The mercapturic acid synthesis is generally a detoxifying pathway which allows elimination of numerous xenobiotics or endogenous compounds as polar metabolites [1]. In the last decade several works pointed out a pro-oxidant role of GGT linked to the metabolism of GSH in the presence of transition metals. This pro-oxidant effect is based on the autoxidation of cysteinylglycine (CysGly), the GGT-generated metabolite of GSH, which produces thiyl and oxygen radicals [2, 3]. Studies performed with different models demonstrated that the GGT-initiated production of reactive oxygen species can result in oxidative damage of various lipid substrates such as polyunsaturated fatty acids [2], human LDL [4] , microsomes [5], as well as oxidative mutagenesis [6] and protein thiolation [7]. Besides, modulatory effects of this pro-oxidant system on transcription factors [8, 9] and cell proliferation [10] have been demonstrated recently. Furthermore, we have previously shown that another GGT-related enzyme (GGT-rel), which is able to metabolise GSH to CysGly, initiates also a prooxidant process leading to peroxidation of membrane lipids [11]. 4-Hydroxy-2,3-/ra«5-nonenal (HNE) is a highly reactive lipid peroxidation (LPO) product which presents cytotoxic and genotoxic effects and interferes with a broad number of cellular functions: gene expression, protein and nucleic acids synthesis, cell proliferation and enzymatic activities [12, 13]. The toxicity of HNE is due to the presence of 3 functional groups: a Cl aldehyde group, a C2=C3 double bond and a hydroxyl group. The C3 position of HNE is a highly reactive site for Michael addition reactions with sulfhydryl groups forming adducts with low molecular mass thiols or proteins [14]. Mammalian cells posses efficient enzymatic equipments which metabolize HNE to less reactive compounds and hence protect proteins against damage induced by this LPO product. The three main pathways of HNE metabolism are: (i) oxidation to innocuous 4-hydroxynonenoic acid (HNA) by aldehyde dehydrogenase, (ii) reduction to unreactive metabolite 1,4-

198

M. Enoiu et al. /Metabolism and Cytotoxicity

dihydroxynonene (DHN) by alcohol dehydrogenase, aldose reductase and several aldo-keto reductases and (in) conjugation with GSH [15-19]. Glutathione S-transferase enzymes catalyze conjugation of GSH to C-3 carbon of HNE, thereby preventing further nucleophilic addition to this toxic compound. On the other hand, although GSH conjugation represents in general a detoxification pathway, there is evidence indicating that the GSH conjugates of a variety of chemicals such as dichlorovinyl [20], bromobenzene, 2-bromohydroquinone [21], polyphenols [22], 17-(5-estradiol [23] and acrolein [24] are further converted to nephrotoxic intermediates. Such an activation requires processing through the GGT-catalyzed reaction which initiates the mercapturic acid pathway and in most of the cases can be prevented by acivicin, a GGT inhibitor. In the particular case of GS-HNE, the mercapturic acid was identified as an end metabolite, together with mercapturic acids of DHN, HNA, and HNA lactone, dicarboxylic acids and their mercapturic acid conjugates [25, 26]. Nevertheless, the biological significance of these metabolites has not been so far studied. We have shown very recently that GGT is able to metabolize GS-HNE to CysGlyHNE and that this enzymatic reaction results in an increased cytotoxicity in a cellular model overexpressing GGT [27]. The present work focuses on the role of GGT in the metabolism of GS-HNE conjugate with special emphasis to the possible lexicological consequences of this metabolic step in amplifying damages induced by (GGT-dependent) lipid peroxidation. We show that GGT-rel recognizes also GS-HNE as a substrate and finally we discuss the new resulting insights in the complex relation between thiol metabolism and oxidative stress phenomena.

2. Materials and Methods 2.1. Chemicals HNE diethyl acetal purchased from OXIS International, Inc. (Portland, OR, USA) was hydrolyzed (0.02 mmol) just prior to use in 1 ml of 1 mM HC1 (Ih, 37°C). Free HNE was extracted by chloroform. After evaporation of the organic layer to dryness, the residue was dissolved in water and HNE concentration was determined spectrophotometrically at 223 nm (8=13750 M-W). GSH, GGT (from bovine kidney), orf/io-phthalaldehyde (OPA) and arachidonic acid were obtained from Sigma Chemical Co. (St. Louis, MO, USA); cysteinylglycine (CysGly) was from Bachem (AG, Bubendorf, Switzerland) and 2-mercaptoethanol (ME) was from Merck (Darmstadt, Germany). All reagents and solvents were of analytical, HPLC or cell culture grade. Solutions were prepared using Milli-Q water (18.2 MQ.cm). 2.2. Synthesis of GS-HNE and CysGly-HNE HNE was incubated with a 10 molar excess of GSH in 50 mM phosphate buffer (pH 7.8) at room temperature for 2 to 3 h. The reaction was monitored by following the decrease of absorbance at 223 nm. The absence of unreacted HNE was also verified using HPLC. GSHNE was purified from GSH excess by solid phase extraction (SPE) through a SEP-PAK Cig cartridge (Waters, Millipore Corporation, Milford, Mass., USA) as described [27]. The adduct of CysGly with HNE was synthesized from GS-HNE by treatment with GGT. Samples containing 200 nmoles GS-HNE were incubated with 250 mU GGT in 50 mM phosphate buffer (pH 7.8) at 25 °C. The conversion of GS-HNE into the corresponding

M. Enoiu et al. /Metabolism and Cytotoxicity

199

CysGly-HNE conjugate was monitored by HPLC and the purification of the final product performed by SPE. Both adducts were stored in aqueous solution under nitrogen at -20°C. 2.3. HPLC Analysis GS-HNE and CysGly-HNE were analyzed after precolumn derivatization of amino group with OPA and ME as follows : to 25 al of sample, 25 ul of OPA-ME derivatization reagent was added, vortex-mixed and injected into the HPLC system after exactly 2 min. The HPLC analysis was performed as described previously [27]. 2.4. Cell Lines V79 Chinese hamster lung fibroblast cells transfected with human GOT cDNA and highly expressing GGT (V79 GGT) and V79 mock-transfected cells, exhibiting non-detectable endogenous GGT activity (V79 Cl) [28] were cultured in RPMI 1640 Gibco medium (Grand Island, NY, USA) supplemented with 5% fetal calf serum (Boehringer Mannheim; Mannheim, Germany), 100 units penicillin, 100 ug streptomycin and 0.25 ug amphotericin B per ml of medium, and 2 mM L-glutamine (Sigma Cell Culture). The murine fibroblast NIH3T3 cell line stably expressing GGT-rel cDNA (3T3/GGT-rel) and NIH3T3 mock transfected cells (3T3) established by Heisterkamp et al. [29] were cultured in DMEM medium (Sigma Cell Culture) added with 10% fetal calf serum and antibiotics as above. Cultures were maintained at 37°C in a humidified atmosphere of 5 % CC»2/95% air. For all experiments, cells were seeded in 24-well plates (Corning Costar Corp.; Cambridge, MA, USA) at 105 cells per well 24 h before experiments, and were then at about 90% confluency. 2.5. Formation ofHNE Adducts during GGT-Dependent Lipid Per oxidation Adherent V79 GGT and V79 Cl control cells were gently washed twice with 2 mL PBS and incubated in PBS (pH 7.4) containing 2 mM GSH, 20 mM glycylglycine (glygly), 150:165 uM Fe3+-EDTA, and 2mM arachidonic acid as a substrate for lipid peroxidation. After 5 hours of incubation at 37°C the medium was purified by SPE and analyzed by HPLC for GS-HNE and CysGly-HNE adducts. 2.6. Metabolism of GS-HNE in Cells in Culture Cells expressing GGT-rel as well as control cell lines were gently washed with PBS and incubated in PBS containing 200 uM GS-HNE at pH 7.4 in a COi incubator. For kinetic measurements of GS-HNE and CysGly-HNE, aliquots of 25 ul medium were taken out at predetermined times and analyzed by HPLC. 2.7. Cytotoxicity Assays For Cytotoxicity assays, medium of 24 h cultured cells was replaced with 400 ul RPMI or DMEM medium containing 0.5% fetal calf serum and supplemented with 50 to 200 uM GS-HNE or CysGly-HNE. Incubations were performed for 3 to 24 h at 37°C in a CO2 incubator.

200

M. Enoiu el al. /Metabolism and Cytotoxicity

Cell viability was measured using MTT and Trypan Blue assays as described previously [27]. In both assays, all individual values were related to the control values (=100%) obtained from untreated cells. Lactate dehydrogenase (LDH) activity released from the cytosol of damaged cells was measured using the LDH cytotoxicity detection kit (Boehringer Mannheim, Mannheim, Germany). Results are expressed as absorbance values obtained after substraction of corresponding untreated controls. 3. Results 3.1. CysGly-HNE is formed during GGT-Dependent Lipid Peroxidation We have shown previously that GS-HNE is a substrate of GGT and the resulting metabolite is CysGly-HNE, as confirmed by mass spectrometry analysis [27]. We have found also that this metabolism takes place in a cell line overexpressing GGT (V79 GGT), but not in GGTnegative control cells, V79 Cl [27]. It is known that in the presence of Fe3+, the metabolism of GSH in V79 GGT cells generates recative oxygen species [3]. Lipid peroxidation is one of the consequences of this GGT-dependent pro-oxidant process, but so far there is no evidence for the production of HNE in this system. Generally, HNE arising from lipid peroxidation is rapidly converted to less reactive metabolites [19]. For this reason, and because of the presence of thiols (GSH and its metabolite, CysGly) in models for studying GGT-induced lipid peroxidation, it is understandable why the formation of HNE in such models is difficult to be demonstrated. Therefore it would be more reliable to investigate the formation of HNE metabolites than that of free aldehyde.

B

Retention time (min) Fig. 1. HPLC detection of thlol-HNE adducts formed during GGT-dependent lipid peroxidation in V79 GGT (A) and V79 Cl (B) cells.

M. Enoiu et al. /Metabolism and Cytotoxicity

201

To find out whether conjugates of HNE could be formed during GGT-dependent lipid peroxidation, we used arachidonic acid as a substrate whose oxidation supplies HNE, and V79 GGT cells in pro-oxidant conditions. After 5 hours incubation of V79 GGT and V79 Cl cells in the presence of GSH, glygly as an acceptor substrate and Fe3+-EDTA, the medium was purified by SPE from unreacted fatty acid and free aldehydes, and analyzed by HPLC. In these conditions, CysGly-HNE could be detected in the medium of V79 GGT cells, GS-HNE being absent or present only in very low amounts (Fig. 1 A). The presence of these adducts was assessed by comparison of HPLC profiles and retention times with those of the corresponding synthetic adducts. In contrast, in the V79 Cl control cells, only the GS-HNE was detected, its presence being very probably due to the autoxidation of the lipid substrate, supplying 4hydroxynonenal, followed by its reaction with GSH (Fig. 1 B). Although the complexity of the chromatograms does not allow an accurate quantification of peaks corresponding to these adducts, the formation of GS-HNE by autoxidation of fatty acid in the control cells can be estimated at less than 50% of the total adducts produced in V79 GGT cells. These results demonstrate that CysGly-HNE conjugate could be formed during GGT-dependent lipid peroxidation. 3.2. GS-HNE is a Substrate ofGGT-rel The GGT-rel enzyme exhibits a substrate specificity different from GGT. Up to now only GSH and leukotriene C4 are known to be substrates for GGT-rel [11, 29]. We aimed at verifying if GGT-rel recognizes also GS-HNE conjugate as a substrate. NIH3T3/GGT-rel cells were incubated in the presence of 200 fxM GS-HNE and the medium analysed by HPLC for GSH adducts. In these conditions NIH3T3/GGT-rel were able to metabolize GSHNE to CysGly-HNE (Fig. 2).

GS-HNE

Retention time (min) Fig. 2. HPLC chromatograms of GS-HNE and CysGly-HNE In the medium of NIH3T3/GGT-rel cells.

202

M. Enoiu et al. /Metabolism and Cytotoxicity

However, the metabolism of GS-HNE in these cells was significantly slower than in V79 GGT cells, which completely metabolized the same amount of conjugate during 2 hours [27]. After 6 hours of incubation we found a CysGly-HNE concentration of 125.5 ± 3.5 uM in the medium of NIH3T3/GGT-rel cells, the remainder 74.4 ± 3.6 uM GS-HNE being still untransformed. In NIH3T3 cells, GS-HNE remained untransformed during the same incubation time and no CysGly-HNE could be detected. 3.3. Cytotoxicity Associated with GGT-rel Metabolism of GS-HNE Our previous data pointed out a cytotoxic effect linked to the metabolism of GS-HNE in V79 GGT cells [27]. In order to find out whether a similar effect accompanies GGT-rel metabolism of GS-HNE, we incubated NIH3T3/GGT-rel and NIH3T3 cells in the presence of different concentrations of GS-HNE.

B

120^

| 80-

*

60

J

40-

201 0 0-

5

GS-HNE (MM)

T

'

*

•

*

50

100

100

GS-HNE (MM)

Fig. 3. Dose dependence of GS-HNE effect on cell viability (A: MTT assay, B: Trypan Blue exclusion assay) In NIH3T3/GGT-rel (•) and NIH3T3 (D) cells. Cells were incubated 24 h in the presence of GS-HNE at the indicated concentrations. Results expressed as % of untreated control are means ± SD of three independent experiments. Statistical significance of differences between NIH3T3/GGT-rel and NIH3T3 cell lines: * p<0.05 ; *** p<0.001.

25

50

75

100

GS-HNE (|iM) Fig. 4. Dose dependence of GS-HNE effect on LDH release in NIH3T3/GGT-rel (•) and NIH3T3 (D) cells. Cells were incubated 24 h in the presence of GS-HNE at the indicated concentrations. Data expressed as absorbance (A«90nm-AWOnJ (average values of corresponding untreated controls are subtracted from each absorbance value) are means ± SD of three independent experiments. Statistical significance of differences between NIH3T3/GGT-rel and NIH3T3 cell lines: * p<0.05 ; ** p<0.01; "* p<0.001.

M. Enoiu et al. /Metabolism and Cytotoxicity

203

After 24 h incubation, the cell viability (measured by MTT and trypan blue exclusion assays) was significantly reduced in NIH3T3/GGT-rel as compared to NIH3T3 cells and 100% cellular death was obtained in NIH3T3/GGT-rel cells at the concentration of 100 uM GS-HNE (Fig. 3). The LDH release, an index of cell death and cell lysis, increased linearly with the concentration of GS-HNE in NIH3T3/GGT-rel cells (Fig.4). Therefore, the three different assays we used demonstrate that the metabolism of GS-HNE by GGT-rel is also accompanied by an increase of cytotoxicity.

4. Discussion The activity of cleavage of gamma-glutamyl moiety confers to GGT enzyme an important role in the metabolism of GSH, GSH conjugates and consequently in the regulation of intracellular GSH level, of cellular redox state and in the pathway of mercapturic acids synthesis, respectively. However, depending on different factors, the GGT-catalyzed reaction could result in contradictory effects at cellular level. Thus the function of GGT in providing amino acids precursors for the intracellular synthesis of GSH confers to this enzyme an antioxidant role. On the contrary, when the initiation of extracellular GSH degradation by GGT occurs in the presence of iron, a pro-oxidant event could take place and oxidative damages such as lipid peroxidation, protein thiolation and oxidative mutagenesis could consequently be induced [2-7]. Conjugation with glutathione is a widely employed mechanism for detoxification of many electrophilic compounds. GSH conjugates are actively extruded out of the cell and transported to the kidney where they are converted to mercapturic acids in a metabolic pathway involving firstly the cleavage of y-glutamyl moiety by GGT. Mercapturic acid synthesis is generally a detoxification pathway. However, a limited number of compounds, including some polyphenols [22, 23], halogenated hydrocarbons [20, 21] and acrolein [24], have been found to be activated to more potent toxins via the mercapturic acid pathway. The metabolite(s) responsible for the enhanced toxicity are either a breakdown product of GSH conjugate (cysteinylglycine or cysteinyl conjugate, mercapturic acids), in some cases more reactive than their precursor, or metabolites formed by further enzymatic activation. Nevertheless, the role of GGT in this metabolic activation seems essential since its inhibition by acivicin can prevent in most of cases the nephrotoxic effect of these GSH conjugates. In the particular case of HNE, conjugation with GSH seems an important mechanism for inactivation of this reactive lipid peroxidation product and its further metabolite HNE-MA has been found in urine [25, 26]. However, the metabolism of GSHNE through the mercapturic acid pathway and the biological significance of the resulting metabolites has not been so far studied. Our work is focused on the GGT-catalyzed step of GS-HNE metabolism. We demonstrated recently that purified GGT transforms GS-HNE to CysGly-HNE, whose identity was confirmed by mass spectrometry. This metabolism was then studied in a cellular model consisting of two cell lines, one overexpressing GGT and another one lacking GGT, allowing to point out the eventual biological effects associated to this enzymatic reaction. As expected, only V79 GGT cells were able to metabolize GSHNE to CysGly-HNE. The major finding of our work was that a cytotoxic effect GGT-, dose- and time-dependent has been found to be associated with the GS-HNE metabolism in V79 GGT cells, whereas in V79 Cl cells no cytotoxic effect was obtained. The observed cytotoxicity could be attributed to the formation of CysGly-HNE. Indeed, treatment of cells

204

M. Enoiu et al. /Metabolism and Cytotoxicity

directly with CysGly-HNE resulted in an identical cytotoxic effect in V79 GOT and V79 Cl cells [27]. These findings are evidence for a cytotoxic effect of CysGly-HNE significantly higher than its precursor, GS-HNE. However, the mechanism of this cytotoxicity is so far speculative. We can hypothesize that the dissociation of CysGly-HNE to HNE and CysGly could be responsible for this effect, since in the incubation conditions we used, but in a cellfree experiment, higher amounts of free HNE were found to be released from CysGly-HNE than from GS-HNE [27]. Another possible mechanism could be based on a direct toxicity of CysGly-HNE which could react with thiol or amino groups of proteins, with formation of a mixed diadduct. Such a reaction seems possible, since we found that in the presence of a thiol excess, the attempt of direct synthesis of CysGly-HNE led to a diadduct [27]. We demonstrated that GGT-rel, another enzyme able to cleave the y-glutamyl moiety, could transform GS-HNE to CysGly-HNE. The physiological role of this GGT homologue protein is unknown and its substrate specificity is different from GGT. GGT-rel was demonstrated to catalyze the cleavage of GSH to CysGly [11] and conversion of leukotriene C4 to leukotriene D4, but it cannot cleave the synthetic chromogenic substrates of GGT [29]. Our results show that GS-HNE is another common substrate of these two enzymes. Moreover, as in the case of GGT, we found that the metabolism of GS-HNE to CysGly-HNE by GGT-rel resulted in an increased cytotoxicity in GGT-rel expressing cells as compared to the control cells lacking GGT-rel. We previously demonstrated that generation of the reactive metabolite CysGly by GGT-rel in the presence of iron could account for a pro-oxidant function of this enzyme [11]. Our present data suggest that GGTrel can also act similarly to GGT in the metabolism of GS-HNE conjugate to a more toxic metabolite, providing evidence for a possible role of both enzymes in this metabolic pathway which is indirectly related to the oxidative stress phenomena. Numerous studies demonstrated that GGT/GSH/Fe3* is an oxidant system able to induce various biological consequences, including lipid peroxidation [2, 4, 5]. In most of cases, this effect was demonstrated by dosage of malonaldehyde, a widely used indicator of lipid peroxidation. However, other aldehydes resulting from lipid peroxidation, in particular 4-hydroxynonenal, are known to have high toxic effects on numerous cell functions, being involved in the aethiology of many human diseases, e.g. atherosclerosis and cancer [12-14]. Although HNE is considered as a better indicator and it was assumed as a model molecule for oxidative stress, its presence has never been signaled in the GGT-dependent lipid peroxidation studies. On the other hand, because of the very rapid metabolism of HNE in cells, it was suggested that the dosage of HNE metabolites would be a more reliable indicator for the generation of HNE than that of free aldehyde. We used our cellular model to check whether the GGT/GSH/Fe3+-mediated lipid peroxidation of a n-6 polyunsaturated fatty acid (arachidonic acid) generates HNE. The presence of the last one was pointed out by the HPLC analysis of its metabolites, GSH-HNE or CysGly-HNE conjugates. In these conditions both adducts could be detected in the medium of cells after incubation. However, the CysGly-HNE was present only in the medium of V79 GGT cells and the relative amount of total thiol-HNE adducts in GGT-expressing cells was about 2 fold higher than the GS-HNE amount found in V79 Cl cells. These results demonstrate that GGT/GSH/Fe3+ increases the level of autoxidation of arachidonic acid to HNE and that its conjugate formed with GSH present in our system is further metabolized by GGT to CysGly-HNE. The results obtained in this cellular model do not allow an extrapolation to in vivo conditions. However, if similar processes could occur in vivo we can then imagine the following scenario: oxidation of membrane polyunsaturated fatty acids to HNE during the metabolism of GSH by GGT in the presence of an iron source could be followed by the conversion of initially formed nontoxic GS-HNE conjugate to the more toxic CysGly-HNE (Fig. 5). This last step could have a particular toxicological significance leading to an

205

M. Enoiu et al. /Metabolism and Cytotoxicitv

enhanced toxicity as a consequence of a LPO process in the microenvironment of GGT-rich cells. That could be the case of kidney cells, which normally express high GGT activity, or in pathological conditions, of tumors known to be strongly GGT-positive (liver, colon, ovary, prostate) [30]. It is also conceivable that GGT-rich cells could be a more distant target for late toxic effects of HNE, when its GSH conjugate reaches these tissues. Extracellular medium

Cytosol

Fe

GSH

LPO

S-HNE <

> GS-HNE

HNE

CysGly-HNE | HNE Cytotoxic effect release ?

Fig. 5. Proposed scheme for amplifying damages due to Hold peroxidation by GGT (GGTrel) activity. The metabolism of extracellular GSH by GGT supplies aminoacids for intracellular de novo GSH synthesis. Alternatively, in the presence of an iron source, autoxidation of CysGly initiates a reactive oxygen species (ROS) production which consequently leads to peroxidation of membranary polyunsaturated fatty acids with formation of HNE. The lipid peroxidation product, HNE, initially detoxified by conjugation with GSH could be again converted by GGT (GGT-rel) to a cytotoxic metabolite. This metabolic activation could occur also for the GS-HNE formed in the intracellular space by glutathione S-transferase (GST) catalysed reaction and exported from the cell, as well as when GS-HNE transported from a more distant LPO site encounters GGT (GGT-

rel).

Structure-activity relationship studies demonstrated that the Cl-aldehyde is the most important functional group in the cytotoxicity of HNE [31]. On the basis of these findings we can hypothesize that as long as the HNE moiety in the conjugate with GSH is not further submitted to oxidoreductive transformations (to HNA or DHN), it keeps its toxic potential and could become more active when encounters GGT. On the other hand, numerous studies pointed out that at physiological concentrations (< 0.1 to 1 fiM) or moderately increased, HNE exerts modulatory effects on several cellular

206

M. Enoiu et al. /Metabolism and Cytotoxicity

functions such as cell proliferation and differentiation, signal transduction, and gene expression. It was therefore suggested that HNE could act not only as a "second toxic messenger" of oxidative stress, but also as a biological signal in normal and pathological conditions [13]. If the delivery of HNE from CysGly-HNE were the mechanism responsible for the GGT-dependent toxic effect of GS-HNE we found, further studies should be performed at lower GS-HNE dosis to test the possibility of modulatory effects similar to HNE itself. Other directions for future investigations could focus on the mechanism of the observed cytotoxicity: identify the molecular entity which is responsible for this effect (CysGly-HNE or HNE) as well as the cellular targets, i.e. cellular membranes or intracellular components. For these purposes, investigation of HNE-protein adducts and transfection of cells with a HNE metabolizing enzyme, such as aldehyde dehydrogenase [31] would provide useful tools. The use of a different cellular model possessing the whole enzymatic equipment of mercapturic acid pathway, such as renal proximal tubular cells [32] would clarify whether the toxic effect we found after the GGT step is relevant or not for the entire metabolic pathway. In any case in vivo studies will be necessary to verify if such an activation could have implications for nephrotoxicity. The last one could be hypothesized by analogy with the nephrotoxic effect of GSH conjugate of acrolein, which is also a reactive aldehyde resulting from lipid peroxidation. Certainly, the implications raised here require further investigations as well as confirmation using different models. However, our results provide the first evidence for a role of GGT (GGT-rel) in the activation of GS-HNE to a cytotoxic metabolite and allow formulation of hypothesis concerning the involvement of this enzymatic activity in enhancement of toxic effects associated to lipid peroxidation.

Acknowledgment This work is a part of PhD thesis of M. Enoiu, performed in collaboration between Centre du Medicament, Universite Henri Poincare Nancy 1 (France) and University Carol Davila Bucharest (Romania).

References [1]

N.P. Curthoys and R.P. Hughey, Characterization

and physiological

function of rat renal y-

glutamyltranspeptidase. Enzyme 24 (1979) 383-403. [2] A.A. Stark, E. Zeiger and D.A. Pagano, Glutathione metabolism by y-glutamyltranspeptidase leads to Hpid peroxidation. characterization of the system and relevance to hepatocarcinogenesis, Carcinogenesis 14 (1993) 183-189. [3] R. Drozdz, C. Parmentier, H. Hacnad, P. Leroy, G. Siest and M. Wellman, f-Glutamyttransferase dependent generation of reactive oxygen species from a glutathioneAransferrin system. Free Radic. Btol. Med. 25 (1998)786-792. [4] A. Paolicchi, G. Minotti, P. Tonarelli, R. Tongiani, D. De Cesare. A. Mezzeti. S. Dominici. M. Comport and A. Pompella, y-Glutamyl transpeptidase-dependent iron reduction and LDL oxidation - a potential mechanism in atheroderosis, J. Invest. Med. 47 (1999) 151-160. [5] A Paolicchi, R. Tongiani, P. Tonarelli, M. Comport and A. Pompella, Gamma-glutamyl transpeptidasedependent lipid peroxidation in isolated hepatocyles and HepG2 hepatoma cells, Free Radic. Bid. Med. 22(1997)853-860. [6] A.A. Stark and G.A. Glass, Role of copper and cerutoplasmin in oxidative mutagenesis induced by the glutathione-rglutamyttranspeptidase system and by other thiols, Environ. Mo). Mutagen 29 (1997) 63-72.

M. Enoiu et al. /Metabolism and Cytotoxicity

207

[7] S. Dominici, M. Valentin!, E. Maellaro, B. Del Bello, A. Paolicchi, E. Lorenzini, R. Tongiani, M. Comport) and A. Pompella, Redox modulation of cell surface protein thiols in U937 lymphoma cells: the role of v-glutamyl transpeptidase-dependent H2O2 production and S-thiolation, Free Radic. Biol. Med. 27 (1999) 623-635. [8] M.-J. Accaoui, M. Enoiu, M. Mergny, C. Masson, S. Dominici, M. Wellman and A. Visvikis, Gammaglutamyltranspeptidase-dependent glutathione catabolism results in activation of NF-kB, Biochem. Biophys. Res. Commun. 276 (2000) 1062-1067. [9] E. Maellaro, S. Dominici, B. Del Belto, M.A. Valentin!, L. Fieri, P. Perego, R. Supino, F. Zunino, E. Lorenzini, A. Paolicchi, M. Comport) and A. Pompelia, Membrane gamma-glutamyl transpeptidase activity of melanoma cells: effects on cellular H2O2 production, cell surface protein thiol oxidation and NF-kappa B activation status, J. Cell. Sci. 113 (2000) 2671-2678. [10] B. Del Bello, A. Paolicchi, M. Comporti, A. Pompella and E. Maellaro, Hydrogen peroxide produced during Y-glutamyl transpeptidase activity is involved in prevention of apoptosis and maintenance of proliferation in U937 cells. FASEB J. 13 (1999) 69-79. [11] M. Enoiu, H. Aberkahne, J.-F. Salazar, P. Leroy, J. Groffen, G. Siest and M. Wellman, Evidence for the pro-oxidant effect of y-glutamyl transpeptidase-related enzyme, Free Radic. Biol. Med. 29 (2000) 825-833. [12] M. Comporti Lipid peroxidation and biogenic aldehydes: from identification of 4-hydroxynonenal to further achievements in biopathology, Free Radic. Res. 28 (1998) 623-635. [13] M. Parola, G. Bellomo, G. Robino, G. Barrera and M. U. Dianzani, 4-Hydroxynonenal as a biological signal: molecular basis and pathophysiological implications, Antiox. Redox Signal. 1(1999) 255-283. [14] H. Esterbauer, R.J. Schaur and H. Zollner, Chemistry and biochemistry of 4-hydroxynonenal, malonaldehyde and related aldehydes, Free Radic. Biol. Med. 11 (1991) 81-128. [15] D.P. Hartley, J.A. Ruth and D.R. Petersen, The hepatocellular metabolism of 4-hydroxynonenal by alcohol dehydrogenase, aldehyde dehydrogenase, and glutathione S-transferase, Arch. Biochem. Biophys. 316 (1995)197-205. [16] S. Srivastava, B.L. Dixit, J. Cai, S. Sharma, H.E. Hurst, A. Bhatnagar and S.K. Srivastava, Metabolism of lipid peroxidation product, 4-hydroxynonenal (HNE) in rat erythrocytes: role of aldose reductase, Free Radic. Biol. Med. 29 (2000) 642-651. [17] M. Burczynski, G. Sridhar, N. Palackal and T. Penning, The reactive-oxygen species- and Michael acceptor-inducible human aldo-keto reductase AKR1C1 reduces the a, /3-unsaturated aldehyde 4hydroxy-2-nonenal to 1,4-dihydroxy-2-nonene, J. Biol. Chem. 276 (2001) 2890-2897. [18] S. Singh, A. Janecki, S. Srivastava, S. Awasthi, Y. Awasthi, S. Xia, S. Xia, and P. Zimniak, Membrane association of glutathione S-transferase mGSTA4-4, an enzyme that metabolizes lipid peroxidation products, J. Biol. Chem. 277 (2002) 4232-4239. [19] P.J.M. Boon, H.S. Marinho, R. Costing and G.J. Mulder, Glutathione conjugation of 4-hydroxy-frans-2,3nonenal in the rat in vivo, the isolated perfused liver and erythrocytes, Toxicol. Appl. Pharmacol. 159 (1999)214-223. [20] Elfarra A.A. and Anders M.W., Renal processing of glutathione conjugates. Role in nephrotoxicity, Biochem. Pharmacol. 33 (1984) 3729-3732. [21] S.S. Lau and T.J. Monks, The in vivo disposition of 2-bromo-[14C]hydroquinone and the effect of Y-glutamyl transpeptidase inhibition, Toxicol. Appl. Pharmacol. 103 (1990) 121-132. [22] T.J. Monks and S.S. Lau, The pharmacology and toxicology of polyphenolic-glutathione conjugates, Annu Rev. Pharmacol. Toxicol. 38 (1998) 229-255. [23] M. Butterworth, S.S. Lau and T.J. Monks Formation of catechol estrogen glutathione conjugates and gamma-glutamyl transpeptidase-dependent nephrotoxicity of 17beta-estradiol in the golden Syrian hamster. Carcinogenesis 18 (1997) 561-567. [24] J.J. Horvath, C.M. Witmer and G. Wftz, Nephrotoxicity of the 1:1 acrolein-glutathione adduct in the rat, Toxicol. Appl. Pharmacol. 117 (1992) 200-207. [25] T. Petras, W.G. Siems and T. Grune, 4-hydroxynonenal is degraded to mercapturic acid conjugate in rat kidney, Free Radic. Biol. Med. 19 (1995) 685-688. [26] J. Alary, F. Bravais, J.P. Cravedi, L. Debrauwer, D. Rao and G. Bories, Mercapturic acid conjugates as urinary end metabolites of the lipid peroxidation product 4-hydroxy-2-nonenal in the rat, Chem. Res. Toxicol. 8(1995)34-39.

208

M. Enoiu et al. /Metabolism and Cytotoxicity

[27] M. Enoiu, R. Herber. R. Wennig. C. Marson, H. Bodaud, P. Leroy, N. Mitrea. G. Siest and M. WeNman, r glutamyttranspeptidase-dependent metabolism of 4-hydroxynonenal-glutathione conjugate. Arch. Biochem. Biophys. 397 (2002) 18-27. [28] A. Visvikis, C. Thbudellet, T. Oster, S. Foumel-Gigleux, M. Wellman and G. Siest, High level expression of enzymatically active mature human gamma-glutamyltransferase in transgenic V79 Chinese hamster cells, Proc. Natl. Acad. Sci. USA 88 (1991) 7361-7365. [29] N. Heisterkamp, E.R. De Meyts, L Uribe, H.J. Forman and J. Groffen, Identification of a human y-glutamy) cleaving enzyme related to, but distinct form, y-glutamyl transpeptidase. Proc. Natl. Acad. Set. USA 88 (1991)6303-6307. [30] M. Hanigan, y-Glutamyltranspeptidase, a glutathionase: its expression and function in cartinogenesis, Chem. Biol. Interact. 111-112 (1998) 333-342. [31] R. Haynes, L Szweda, K. Pickin, M. Welker, and A. Townsend, Structure-activity relationships for growth inhibition and induction of apoptosis by 4-hydroxy-2-nonenal in Raw 264.7 cells, Md. Pharmacol. 58 (2000) 788-794. [32] P.J. Boogard, J.F. Nagelkerde, and G.J. Mulder, Renal proximal tubular cells in suspension or in primary culture as in vitro models to study nephrotoxicity, Chem. Bid. Interact. 76 (1990) 251-291.

Thiol Metabolism and Redox Regulation of Cellular Functions A. Pompella etui (Eds.) IOS Prcsx, 2002

209

y-Glutamyltransferase-DependentProoxidant Reactions: a Factor in Multiple Processes Silvia DOMINICI, Aldo PAOLICCHI, Evelina LORENZINI, Emilia MAELLARO*, Mario COMPORTI*, Lisa FIERI, Giorgio MINOTTI§ and Alfonso POMPELLA Dept. of Experimental Pathology, University of Pisa Medical School, (f) Dept. ofPathophysiology and Experimental Medicine, University of Siena, and (^)Dept. of Drug Sciences, G. D 'Annunzio University Medical School, Chieti, Italy

1. Introduction Evidence has accumulated over the past decade that number of cellular functions can be modulated by prooxidants - e.g., reactive oxygen species (ROS) - at concentrations considerably lower than those capable to induce oxidative injury. Prooxidants thus can no longer be regarded as merely offensive species, and similarly, the physiological role of some established 'antioxidants' also is in need of careful reconsideration. Glutathione (GSH) - perhaps the best known cellular antioxidant - appears an ideal candidate in this perspective. The antioxidant role of GSH is readily apparent in detoxification of electrophilic/oxidizing drugs and protection from lipid peroxidation. Nevertheless, nonantioxidant functions of GSH have been decribed e.g. in modulation cell proliferation, immune response and neurotransmission. To complete the picture, recent studies point toprooxidant effects of (extracellular) GSH, which can ensue from its catabolism by the membrane ecto-enzyme gamma-glutamyltransferase (GGT). It has in fact been documented in our and other laboratories that prooxidant species (superoxide, HiOi, thiyl radicals) are produced during GSH catabolism, as a result of the interaction of GSH metabolites - cysteinyl-glycine in the first place - with trace levels of iron ions present hi the cell environment. These phenomena appear to outline an additional, as yet poorly appreciated function of GGT in the modulation of cell redox status, which adds to its well-established role in the cellular metabolism of glutathione and sulfur aminoacids. 2. y-GIutamyltransferase-Dependent Generation of ROS and Other Free Radicals Gamma-glutamyltransferase activity (EC 2.3.2.2), normally found in serum as well as in the plasma membrane of virtually all cells, catalyzes the first step in the degradation of extracellular GSH, i.e. the hydrolysis of the gamma-glutamyl bond between glutamate and cysteine [1]. In so doing GGT releases cysteinyl-glycine, which is subsequently cleaved to cysteine and glycine by plasma membrane dipeptidase activities. Stark et al. [2] first proposed that the catabolism of GSH can play a prooxidant role in selected conditions. These authors suggested that the GGT-mediated cleavage of GSH allegedly through the generation of the more reactive thiol glycyl-cysteine - could cause

210

S. Dominici et al. / y-Glutamyltransferase-Dependent Prooxidant Reactions

the reduction of ferric iron Fe(III) to ferrous Fe(II), thus starting an iron redox-cycling process liable to result in the production of reactive oxygen species and stimulation of oxidative reactions. GOT was thus shown to stimulate a GSH-dependent lipid peroxidation in model systems involving Fe(HI) complexes as redox catalysts and purified linoleic acid peroxidizable substrate [2]. In these systems, the "pro-oxidant" effect of GGT was attributed to the formation of cysteinyl glycine and cysteine, which reduce Fe(III) more efficiently than does GSH [3,4]. Redox cycling of iron in fact leads to the production of ROS [5] - i.e. superoxide anion in the first place, from which in turn by dismutation hydrogen peroxide is originated - as well as of thiyl (-S*) radicals; the latter can easily react to form disulfides. The following overall sequence can be envisaged as set into motion by GGT-mediated catabolism of GSH [2] (GC-SH = cy steiny 1-glycine): GSH

GG7

> glutamicacid + GC-SH;

<1>

GC-SH —pH>7.0- > GC-S- (thiolate anion) + H + ;

<2>

GC-S- + Fe3*

<3>

f GC-S' (thiyl radical) + Fe2*;

Fe2* + Qj

> Fe3* + 02~ (superoxide anion);

<4>

Oj- + H20

> 02 + H202 (hydrogen peroxide).

<5>

The production of ROS following iron reduction induced by the GGT-mediated catabolism of GSH has been repeatedly documented [6-9]. Fig. 1 reports data obtained with U937 histiocytic lymphoma cells, possessing «15 mU GGT/mg protein at their surface. The addition of GSH and co-substrate glycyl-glycine to cells results in a sustained production of hydrogen peroxide. The reaction occurs in the extracellular environment, as shown by the fact that catalase - which cannot penetrate cell membrane - can suppress it (Fig. 1A). Generation of H2O2 did not take place with cells in which GGT had been irreversibly inhibited by the non-competitive GGT inhibitor acivicin, nor in the presence of the competitive GGT inhibitor, serine-borate complex (SBC). H2Ch production was also inhibited in the presence of the vitamin E analogue Trolox C, and was suppressed by addition of low concentrations of the iron chelator deferoxamine (DFO) [6], thus confirming the involvement of extracellular iron ions in the reaction. As expected, H2C«2 production could also be started by the addition to assay mixture of purified GGT protein, as well as by direct addition of cy steiny 1-glycine, i.e. the metabolite resulting from GGT-mediated cleavage of GSH (Fig. IB) [6].

3. Molecular Targets of GSH/GGT-Dependent Prooxidant Reactions The possibility that GSH/GGT-dependent prooxidants could interfere with the redox status of thiol groups contained in proteins of the cell surface was investigated. Conceivably, a primary target for the action of prooxidants generated extracellularly during GGT activity would be given by thiols of proteins located at the cell surface. Therefore, we developed and validated a procedure for the selective labeling of thiols of cell surface proteins [6]. Indeed, the results obtained revealed the occurrence of a GGT-

211

S. Dominici et al. / y-Glutamyltransferase-Dependent Prooxidant Reactions

dependent oxidation of thiol groups in surface proteins of U937 cells. Protein thiol oxidation was in fact increased following stimulation of GOT activity, while the process was prevented after its inhibition. GSH •*• gly-ely U937 cells

buffer buffer -

U937 ceils + acivicin U937 cells + SBCU937 cells + Trolox C -

buffer -

U937cells-

buffer -

HA

U937 cells -

0.5 uM

30

i

60

0

30

\ 60

Time of incubation (min)

Fig. 1. GSH- and GGT-dependent extracellular production of hydrogen peroxide from U937 cells. Decrease of scopoletin fluorescence in the presence of horseradish peroxidase. Vertical bar indicates the fluorescence decrease corresponding to a concentration of 0.5 juM H2O2, as established in preliminary calibration experiments. (A) Production of H2O2 by U937 cells (3 x 106/ml) upon addition of the substrate GSH (100 pM) and the y-glutamyl acceptor glycyl-glycine (gly-gly, 1 mM). H2O2 production was dependent on the extracellular availability of iron ions, as shown by the inhibition offered by adding the iron chelator deferoxamine (DFO, 50 fjM) or the structurally unrelated metal chelator EDTA (50 jvM; not shown). Catalase was directly added to the incubation mixture (100 //g/ml, final concentration). Where indicated, cells were pretreated with the non-competitive GGT inhibitor acivicin (130 /vM, 4 h) or monitored in the presence of the competitive GGT-inhibitor serine/boric acid complex (SBC, 10/20 mM). H2O2 production was also prevented in the presence of the free radical scavenger Trolox C (1 mM). (B) Generation of H2O2 induced by addition of purified GGT protein (corresponding to 18 mU/ml enzyme activity) or cysteinylglycine (10 pM, final concentration), the product of GGT-mediated GSH metabolism. Data from ref. 6, modified.

212

5. Dominiciet ai / y-Glutamyltransferase-Dependent Prooxidant Reactions

Rg. 2. Increase of cell surface protein thiols following Irreversible Inhibition of GGT activity by aclvteln. Confocal imaging of fluorescence levels detected in U937 cells labeled for cell surface protein thiols with the thiol-specific probe maleimidyl-propionyt biocytin / eaFITC (see ref. 10 for details of the procedure). Imaging parameters were adjusted in order to minimize visualization of control cells (panel A), thus making the increase of fluorescence in GGT-inhibited cells immediately apparent (panel B).

GGT inhibition was

obtained

by

exposing U937 cells to 130 pM acivicin for 24 hrs prior to the experiment. Bar corresponds to 10//m.

The involvement of hydrogen peroxide in the process is indicated by the fact that protein thiol oxidation was significantly prevented by catalase. Experiments also showed that GGT-dependent decrease in reduced protein thiols was partly due to protein S-thiolation reactions, and that GGT inhibition by acivicin is per se sufficient to produce an increase of reduced protein thiols at the cell surface (Fig. 2). The latter observation appears to imply that in GGT-rich cells surface proteins are continuously exposed to a GGT-dependent oxidant stress, which maintains their thiols partially oxidized. One of the physiological roles of GSH catabolism by GGT could thus lie in its ability to modulate the redox status of cell surface protein thiols. Subsequent studies were aimed to identify specific macromolecules involved by GGT/GSH-dependent prooxidant reactions. Among several redox-sensitive targets, the transcription factor NF-kB is perhaps the best known and studied [11]. Studies were thus carried out to verify the likely involvement of NF-kB in redox changes consequent to GSH catabolism. Using murine V79-GGT cells, highly expressing a human GGT transgene, it was indeed shown that GGT-dependent ROS production induces the NFkB-binding and transactivation activities. This induction mimicked the one observed by H2O2 and was inhibited by catalase, suggesting the involvement of GSH/GGT-derived H2O2 in the NF-kB activation [12]. However, studies carried out in human tumour cells showed that GSH/GGTdependent modulation of NF-kB activation status can be more complex that firstly appreciated. Stimulation or inhibition of GGT activity in human melanoma Me665/2/60 cells resulted in stimulation or inhibition of NF-kB nuclear translocation, respectively [9]. The increased nuclear translocation following stimulation of GGT activity by the substrates glutathione and glycyl-glycine was however paradoxically accompanied by decreased NF-KB DNA binding and gene transactivation (Fig. 4). NF-KB DNA binding could be restored by treating cell lysates with the thiol-reducing agent dithiothreitol (not

5. Dominid el at. / y-Glutamyltransferase-Dependent Prooxidant Reactions

213

shown), indicating the involvement of critical thiol groups by GSH/GGT-generated prooxidants. These observations indicate that reactions ensuing from GSH catabolism while facilitating mobilization of NF-kB from cytoplasm to the nucleus - can ultimately down-regulate NF-KB DNA-binding and transcriptional activity, thus likely representing a mechanism for preventing excess NF-KB activation in conditions of persistent oxidative stress. Besides NF-kB, modulatory effects by prooxidants and/or antioxidants have also been reported for AP-1 [13]. Interestingly, c-Jun DNA binding activity was recently shown to be redox-regulated through the reversible S-thiolation of a critical cysteine residue [14]; the same authors reported that the phenomenon could also involve the p50 component of NF-kB [15]. Studies in our laboratory have shown that the activation of GSH catabolism through GGT results in increased AP-1 DNA binding. Consistently, the GGT inhibitor acivicin suppressed this effect, confirming the role of GGT-mediated GSH catabolism. AP-1 DNA binding was suppressed also by the independent GGT inhibitor azaserine, as well as by catalase, Trolox C and deferoxamine, confirming that the effect is mediated through GGT-dependent, iron-catalyzed, oxidative mechanisms [16].

4. Effects on Cell Proliferation and Apoptosis It is widely recognized that prooxidants can play a modulatory role on the transduction of proliferative/apoptotic signals, due to their ability to interact with redox-sensitive regions of growth factor receptors, protein kinases and transcription factors [17-19]. A first indication that prooxidant reactions originating from GSH catabolism could play a role in these processes came from studies with human A2780 ovarian cancer cells, showing that exogenous GSH exerts an antiproliferative action, and that this is an effect of HaO2 and thiol oxidation produced by GGT-mediated extracellular GSH catabolism. The antiproliferative effect of GSH in fact was reversed by catalase and by dithiothreitol, indicating the occurrence of oxidative phenomena resulting in the impairment of critical cellular thiols. Treatment of cells with hydrogen peroxide also resulted in growth inhibition in A2780 cells. The Y-glutamyl acceptor glycyl-glycine, a cofactor for GGT activity, potentiated the growth-inhibitory effect of GSH, which in contrast was decreased by the GGT inhibitors, serine/boric acid complex and acivicin, indicating that the production of reactive forms of oxygen, hydrogen peroxide in the fist place, was mediated by glycyl-cysteine produced during GGT-mediated GSH hydrolysis [20]. To remark the complexity of the picture, however, subsequent studies in U937 histiocytic lymphoma cells also showed that a continuous GGT-dependent production of H2O2 can provide tumor cells with a basal, "anti-apoptotic" signal. Previous work had shown that mild oxidative conditions can counteract apoptotic stimuli [8]. Since the inhibition of GGT is a sufficient stimulus for the induction of apoptosis in selected cell lines, we investigated whether this effect might result from the suppression of the mentioned GGT-dependent prooxidant reactions, in the hypothesis that the latter may represent a basal antiapoptotic and proliferative signal for the cell. Experiments with U937 cells showed that: ii) GGT inhibition results in cell growth arrest, and induces cell death and DNA fragmentation with the ladder appearance of apoptosis; iii) treatment of cells with catalase is able to decrease their proliferative rate; iv) GGT inhibition (with

214

5. Dominici el al. / Y-Glutamyltransferase-Dependent Prooxidant Reactions

suppression of FhCh production) results in a down-regulation of poly(ADP-ribose) polymerase (PARP) activity, soon after followed by the proteolytic cleavage of PARP molecule itself, such as that typically induced by caspases [8]. In conclusion, data indicate that the low H2O2 levels originating as a by-product during GGT activity are capable to act as sort of a 'life signal' in U937 cells, insofar as they can maintain cell proliferation and protect against apoptosis, possibly through an up-regulation of PARP activity [8].

5. GSH/GGT-Dependent Lipid Peroxidation Redox cycling of iron is a recognized factor in initiation of lipid peroxidation [21]. Accordingly, GSH/GGT-dependent iron reduction was repeatedly shown to result in the promotion of lipid peroxidation, in several distinct experimental models. Stark et al. [2] first reported on the occurrence of a GGT/GSH-dependent lipid peroxidation in vitro, in systems including Fe(III) complexes and purified linoleic acid. Subsequent studies evidenced the prooxidant action potentially played by GGT reexpressed in chemicallyinduced preneoplastic lesions of rat liver. When fresh cryostat sections of liver were exposed to GSH, chelated Fe(III) and glycyl-glycine, the GGT activity present in transformed cells was able to catalyze the initiation of a lipid peroxidation process, which could be revealed by means of histochemical reactions [22, 23]. Experiments showed that transferrin could serve as a source of redox-active iron. Fig. 3 reports the typical results obtainable in these systems, using the NAH-FBB histochemical reaction for tissue carbonyls developed in our laboratory [24].

Fig. 3. Elective Involvement of GGT-poslttve hepatocytes In GGT-dependent llpM peroxidation. Serial cryostat sections (unfixed) obtained from the liver of a diethylnrtrosamine / 2-acetyl-aminofluorene-treated rat, sacrificed at the end of the initiationpromotion hepatocarcinogenic schedule. (A) GGT activity; (B) Lipid peroxidation (NAH-FBB histochemical reaction); prior to staining the section was incubated (60 min) in the presence of reduced GSH (100 pM) and human h-transferrin (100 /;M); (C) Same treatment as in (B), with addition of glycyl-glycine (1 mM) as acceptor of the GGT-mediated transpeptidation reaction; the area involved by lipid peroxidation has considerably increased as compared to incubation with GSH alone.

5. Dominici el ul. / y-Glutamyltransferase-Dependent Prooxidant Reactions

215

Subsequent studies were extended to living isolated hepatocytes, in which a significant - though non-toxic - lipid peroxidation developed upon exposure to purified GOT, GSH and chelated Fe(III) [25]. The same was observed to occur in HepG2 hepatoblastoma cells, whose significant intrinsic GGT activity was able to directly catalyze the initiation of lipid peroxidation. In rat liver microsomes, the development of GGT/GSH-dependent lipid peroxidation was shown to result in a concomitant oxidation of protein -SH groups [25]. GGT-dependent lipid peroxidation was also observed with isolated human plasma LDL lipoproteins, as detailed below.

6. GSH/GGT-Mediated LDL Oxidation, Atherosclerosis and Progression of Cardiovascular Disease Low density lipoprotein (LDL) oxidation is thought to play a central role in atherogenesis and vascular damage. Iron is a potential catalyst of LDL oxididation, provided that electron donors convert Fe(III) to redox-active Fe(II) (REF). Thiol compounds such as cysteine and homocysteine are known to reduce Fe(III) and promote Fe(II)-dependent LDL oxidation (26). Preliminary histochemical studies had shown that intense GGT activity is detecteble in the intimal layers of human atherosclerotic lesions, where it is apparently expressed by CD68+ macrophage-derived foam cells [27] (Fig. 4). Moreover. GGT-positive foam cells were found to co-localize with immunoreactive oxidized LDL, suggesting a possible role for GGT in the cellular processes of ironmediated LDL damage. Interestingly, catalytically active GGT was also demonstrated in correspondence of microthrombi adhering to the surface of atheromas (Fig. 5). Subsequent studies were thus dedicated to verify the possibility that GGTmediated production of cysteinyl-glycine during GSH catabolism might serve as a mechanism to promote iron reduction and hence LDL peroxidation, thus representing a potential mechanism in progression of atherosclerosis. Experiments showed that in systems including ADP-Fe(III) complexes GSH itself can reduce some iron, but the reaction rate increases significantly when GGT is included to remove its gammaglutamate residue. This effect was observed over a broad range of GGT activities and GSH concentrations [27]. Previous studies had shown that the pKa of the cysteinylglycine thiol is significantly lower than that of GSH [4, 28]. The ability of GGT to enhance iron reduction by GSH might therefore reflect the formation of cysteinylglycine, bearing a thiol moiety which dissociates more rapidly at near-neutral pH, and can thus redox-couple with Fe(III). In agreement with this possibility, cysteinyl-glycine was found to reduce ADP-chelated Fe(III) more effectively than did GSH, forming Fe(II) to the same extent as observed with GSH plus GGT. Similar results were obtained with cysteine, that is the product of cysteinyl-glycine hydrolysis by membranebound dipeptidase. In keeping with GGT/GSH-dependent iron reduction, purified GGT was also found to stimulate GSH/iron-dependent LDL oxidation (Table 1), thus confirming its potential role in pathogenesis of atherosclerosis. Importantly, additional experiments indicated the ability of GGT-mediated GSH catabolism to stimulate the reductive delocalization of iron ions bound to transferrin, i.e. a physiological source of iron (Table 2); in so doing, GGT generates a pool of Fe(II) which readily catalyzes LDL oxidation, as evidenced by the formation of thiobarbituric acid-reactive substances (TEARS). Corollary experiments showed that GSH-dependent LDL oxidation can be efficiently

216

S. Dominici et al. / y-Glutamyltransferase-Dependent Prooxidant Reactions

promoted by cells expressing significant levels of GOT activity at their surface, such as HepG2 hepatoblastoma or U937 monoblastoid cells (Table 3).

Fig. 4. Colocalizatlon (arrow*) of enzymatlcally active GGT with cellular elements of macrophaglc lineage In a human atherosclerotic plaque. (A) Histochemical azocoupling reaction; (B) mouse monoclonal to macrophagtc CD68, stained by an ABC-phosphatase system. Serial cryostat sections from a specimen of coronary artery, obtained at autopsy from a patient deceased 6 hours earlier after pulmonary embolism.

Fig. 5. Enzymatlcally active GGT Is In thrombi adhering to a human atherosclerotic plaque. (A) Haematoxylin/eosin; (B) azocoupling reaction for GGT activity; (C) same as in (B). but incubated in the presence of the GGT competitive inhibitor, serineAwric acid complex.

S. Dominici et al. / Y-Glutamyltransferase-Dependent Prooxidant Reactions

Table 1. Effects of GGT inhibitors on enzyme activity, ADP-Fe(lll) reduction and LDL oxidation.

System

GGT activity

ADP-Fe(lll) reduction a> (nmol Fe(ll) / min)

LDL oxidation b) (nmol TBARS / mg protein)

-

1.4

31

GSH + GGT

200

6.8

83

GSH + GGT + acivicin

1.4

1.9

32

GSH + GGT + SBC

2.1

1.4

34

(mil /ml)

GSH

a

) Incubations (1 ml final volume) contained glycyl-glycine (20 mM), bathophenan-throline

disulfonate (0.25 mM) and ADP-Fe(lll) (1 mM chelator - 100 pM FeCI3). Reactions were started by adding GSH (2 mM) put or minus GGT (200 mU).

b

) Incubations were prepared as in a), with the

exception that bathophenanthroline was omitted and LDL (0.1 mg protein/ml) was included for TBARS assay. Values were determined at 1 h. Where indicated, incubations also contained acivicin (1 mM) or a serine-borate complex (SBC; 10 and 20 mM, respectively). Values are from representative experiments.

Table 2. GSH/GGT-dependent release of transferrln-bound Iron and promotion of LDL oxidation.

Transferrin

Release ofFe(ll), nmol /min a)

(nmol TBARS / mg protein) b)

GSH

ND

ND

GSH + GGT

0.1

5.0

GSH

0.1

8.1

GSH + GGT

0.9

47.2

GSH -H GGT + SOD

1.4

n.d

GSH + GGT + acivicin

0.2

n.d.

GSH + GGT + SBC

0.1

n.d.

Partially saturated

Holo-saturated

In

a

) the incubations (1 ml final volume) were prepared and assayed for bathophenanthroline-

chelatable Fe(ll), released in the presence of partially- or holo-saturated transferrin (100 yM Fe(lll), corresponding to 13.9 or 4.6 mg protein, respectively). Where indicated the incubations also contained acivicin (1 mM), serine-borate complex (SBC, 10-20 mM, respectively) or SOD (200 U). In b) bathophenanthroline was omitted, LDL (0.1 mg protein/ml) was included assayed after 60 min. ND, not detectable; n.d., not determined.

and TBARS were

217

218

5. Dominici el al. / y-Glutamyllransferase-Dependent Prooxidant Reactions

Table 3. Stimulation of GSH- and Iron- dependent LDL oxidation by cell-bound GGT. nmolTBARS/mla)

System

GGT dependent LDL oxidation (nmolTBARS/mg protein) b)

HepG2 cells LDL

3.6 .

LDL+ HepG2 cells

2.0

22.7

171

U937 cells

nd

LDL

nd

LDL+ U937 cells

4.6

46

LDL+ U937 cells+ acivicin

0.5

5

Incubations (1 ml final volume) contained GSH (2 mM), AOP-Fe(lll) (10:1 ratio of chelator to iron) and gtycyl-glyctne (20 mM) in Krebs-Ringer, pH 7.4, 37*C. Where indicated the system also inducted LDL (0.1 mg protein) plus or minus subconfluent HepG2 eel monolayers or U937 cells, in the presence or absence of acivicin (1 mM). Final iron concentration was 150 or 15 (M with HepG2 or U937 ceHs, respectively. In (a) TBARS were assayed in cell-free 1 ml supematants or in 1 mi-suspended HepG2 monolayers or U937 pellets. In (b) net GGT dependent LDL oxidation was determined by correction for background TBARS formation and normalized to mg protein.

Values are taken from representative

experiments.

In conclusion, biochemical and histochemical findings appear to highlight a novel mechanism of LDL oxidation and vascular damage. These observations are of particular interest in consideration of the fact that epidemiologic evidence has been provided for a possible correlation between persistent elevation of serum GGT and mortality from ischemic heart disease [27, 29], a finding that we could recently confirm in a highly controlled clinical study of coronary artery disease patients [30; Emdin et al., this volume]. 7. GGT-Mediated Oxidative Stress during Kidney Ischemia The mechanisms responsible for tubular damage and cell death following acute renal ischemia are still not fully understood. It is accepted that a loss of membrane selective permeability and the collapse of the ionic gradients through cell membrane - as well as the abnormal activation of phospholipases and proteases due to impaired calcium homeostasis - play the main role in the genesis of the biochemical and morphological alterations observed in the proximal tubules during prolonged renal ischemia [31, 32]. Some evidence however exists that oxidative stress phenomena may concur to the pathogenesis of this condition. In fact, despite the very low oxygen concentrations that can be reach in ischemic conditions, free radicals-mediated reactions take place in the ischemic organ and lead to oxidative damage, or at least this may be assumed from the increased lipid peroxidation and/or antioxidant consumption observed in such conditions [33, 34].

S. Dominici et al. / j-Glutamyltransferase-Dependent Prooxidant Reactions

219

In the kidney, GOT is highly expressed expressed on the outer surface of the microvillus membrane (brush border) in the proximal tubule [35]. The potential role of GOT activity - as a promoter of oxidative stress and lipid peroxidation - in the pathogenesis of ischemic tubular cell damage was thus investigated. A rat model of unilateral renal ischemia was set up and the degree of tubular cell damage and lipid peroxidation evaluated. GGT activity was found to be remarkably increased in both cortical and medullar zones of the ischemic kidneys, where GSH levels were only slightly decreased and lipid peroxidation on the contrary was increased (Fig. 6). In parallel, the cytosolic volume of the proximal tubular cells showed a significant increment. The pretreatment of animals with the GGT inhibitor acivicin, besides preventing the up-regulation of the enzyme during ischemia, afforded good protection against the observed lipid peroxidation and changes of GSH levels, as well as of tubular cell volume.

control acivicin + control

ischemia acivicin + ischemia Hi ** 0

1

2

3

4

GGT activity (U/mg protein)

5

2

4

6

MDA (nmol/mg protein)

Fig. 6. Effects of 25 mln of Ischemia on (A) GGT activity and (B) lipid peroxidation (content of malonaldehyde, MDA) of rat kidney cortex. Values are means ± SE from five rats per experimental group. (*) Significantly different from control (P < .05); (**) significantly different from ischemia (P < .001). Data from ref. 36, modified.

Thus, a net pro-oxidant effect of GGT up-regulated during short term ischemia of rat kidney was observed. Up-regulation of GGT appears to contribute to the renal morphological damage exerted by a brief hypoxic condition at the level of proximal tubular cells. Further investigation is required to elucidate the mechanism(s) underlying these phenomena. 8. Concluding Remarks The experimental evidence obtained in our and other laboratories allows to describe a novel aspect of glutathione metabolism, i.e. the redox interactions involving its catabolites originating from its cleavage by GGT. With the mediation of iron - and conceivably of other transition metals as well - GSH catabolism leads to the generation

220

5. Dominici et al. / y-Glutamyltransferase-Dependent Prooxidant Reactions

of ROS and thiyl radicals, whose prooxidant action is detectable on protein thiol groups in the first place. Such processes appear to involve a variety of cellular targets, including important elements of the signal trasduction chains. As GSH/GGT-dependent prooxidant reactions have been described by now in several distinct conditions and experimental models, these phenomena appear to represent a general pathophysiological process, with possible bearings on the understanding and treatment of several human disease conditions.

Acknowledgments The authors are indebted to the Associazione Italiana Ricerca sul Cancro (A.I.R.C., Milan, Italy) for its generous financial support to the research presented in this report.

References 1. Whitfield JB. Gamma glutamyl transferase. Crit. Rev. Clin. Lab. Sti. 38263-355,2001. 2. Stark AA, Zeiger E, Pagano DA, Glutathione metabolism by y-gkitamyl transpeptidase leads to lipid peroxidation: characterization of the system and relevance to hepatocarcinogenesis. Carcinogenesis 14:183-189, 1993. 3. Tien M, Bucher JR, Aust SD, Thiol-dependent lipid peroxidation. Biochem. Biophys. Res. Commun. 107:279-285,1982. 4. Spear N, Aust SD, Thiol-mediated NTA-Fe(lll) reduction and lipid peroxidation. Arch. Biochem. Biophys. 312:198-202,1994. 5. Aust SD, Morehouse LA, Thomas CE, Role of metals in oxygen radical reactions. Free Rad. Biol. Med. 1:3-25,1985. 6. Dominici S, Valentin! M, Maellaro E, Del Belto B, Paolicchi A, Lorenzini E, Tongiani R, Comporti M, Pompella A. Redox modulation of cell surface protein thiols in U937 lymphoma cells: the role of y glutamyl transpeptidase-dependent H202 production and S-thiolation. Free Rad. Bid Med. 27: 623635,1999. 7. Drozdz R, Parmentier C, Hachad H, Leroy P, Siest G, Wellman M. gamma-Glutamyltransferase dependent generation of reactive oxygen species from aglutathione/transferrin system. Free Rad. Biol. Med. 25:786-92, 1998 8. Del Belto B, Paolicchi A, Comporti M, Pompella A, Maellaro E. Hydrogen peroxide produced during gamma-glutamyl transpeptidase activity is involved in prevention of apoptosis and maintainance of proliferation in U937 cells. FASEB J. 13:69-79,1999. 9. Maellaro E, Dominici S, Del Belto B, Valentin! MA, Pieri L, Perego P, Supino R, Zunino F, Lorenzini E, Paolicchi A, Comporti M, Pompella A, Membrane gamma-glutamyl transpeptidase activity of melanoma cells: effects on cellular H2O2 production, cell surface protein thiol oxidation and NF-kB activation status. J. Cell Sci. 113: 2671-2678, 2000. 10. Pompella A, Cambiaggi C, Dominici S, Paolicchi A, Tongiani R, Comporti M. Single-cell investigation by laser scanning confocal microscopy of cytochemical alterations resulting from extracellular oxidant challenge. Histochem. Cell Biol. 105:173-178.1996. 11. Flohe L, Brigelius-Ftohg R, Saltou C, Traber MG, Packer L. Redox regulation of NF-kappa B activation. Free Rad. Biol. Med. 22:1115-1126,1997. 12. Accaoui MJ, Enoiu M, Mergny M, Masson C, Dominici S, Wellman M and Visvikis A, Gammaglutamyltranspeptidase-dependent glutathtone catabolism results in activation of NF-kB. Biochem. Biophys. Res. Commun. 276:1062-1067, 2000.

S. Dominici et al. / y-Glutamyltransferase-Dependent Prooxidant Reactions

221

13. Meyer M, Schreck R, Baeuerle PA. H2OZ and antioxidants have opposite effects on activation of NFkappaB and AP-1 in intact cells: AP-1 as a secondary antioxidant-responsive factor. EMBO J. 12: 2005-2015, 1993. 14. Klatt P, Molina EP, De Lacoba MG, Padilla CA, Martinez-Galesteo E, Barcena JA, Lamas S, Redox regulation of c-Jun DMA binding by reversible S-glutathiolation. FASEB J. 13:1481-1490,1999. 15. Klatt P, Lamas S, Regulation of protein function by S-glutathiolation in response to oxidative and nitrosative stress. Eur. J. Biochem. 267:4928-4944, 2000. 16. Paolicchi A, Dominici S, Fieri L, Maellaro E, Pompella A. Glutathione catabolism as a signalling mechanism. Biochem. Pharmacol. 2002 (in press) 17. Monteiro HP, Stern A, Redox modulation of tyrosine phosphorylation-dependent signal transduction pathways. Free Rad. Biol. Med. 21:323-333, 1996. 18. Lander HM. An essential role for free radicals and derived species in signal transduction. FASEB J. 11:118-124, 1997. 19. Sen ChK. Redox signaling and the emerging therapeutic potential of thiol antioxidants. Biochem. Pharmacol. 55:1747-1758,1998. 20. Perego P, Paolicchi A, Pompella A, Carenini N, Romanelli S , Zunino F. The cell-specific antiproliferative effect of reduced glutathione is mediated by gamma-glutamyl transpeptidasedependent extracelllular prooxidant reactions. Int. J. Cancer 71: 246-250, 1997. 21. Minotti G. Sources and role of iron in lipid peroxidation. Chem. Res. Toxicol. 6:134-46,1993. 22. Pompella A, Paolicchi A, Dominici S, Comporti M, Tongiani R, Selective colocalization of lipid peroxidation and protein thiol loss in chemically induced hepatic preneoplastic lesions: the role of yglutamyl transpeptidase activity. Histochemistry Cell Biol. 106:275-282,1996. 23. Stark A-A, Russell JJ, Langenbach R, Pagano DA, Zeiger E, Huberman E. Localization of oxidative damage by a g!utathione-y-glutamyl transpeptidase system in preneoplastic lesions in sections of livers from carcinogen-treated rats. Carcinogenesis 15:343-348,1994. 24. Pompelta A, Comporti M. The use of 3-hydroxy-2-naphthoic acid hydrazide and Fast Blue B for the histochemical detection of lipid peroxidation in animal tissues - a microphotometric study. Histochemistry 95:255-262, 1991. 25. Paolicchi A, Tongiani R, Tonarelli P, Comporti M, Pompella A. Gamma-glutamyl transpeptidasedependent lipid peroxidation in isolated hepatocytes and HepG2 hepatoma cells. Free Rad. Biol. Med. 22:853-860,1997. 26. Berliner JA, Heinecke JW. The role of oxidized lipoproteins in atherogenesis. Free Rad. Biol. Med. 20:707-727, 1996. 27. Paolicchi A, Minotti G, Tonarelli P, Tongiani R, De Cesare D, Mezzetti A, Dominici S, Comporti M, Pompella A. Gamma-glutamyl transpeptidase-dependent iron reduction and low density Itpoprotein oxidation - a potential mechanism in atherosclerosis. J. Invest. Med. 47:151-160,1999. 28. Stark A-A, Pagano DA, Arad A, Siskindovitch S, Zeiger E, Effect of pH on mutagenesis by thiols in Salmonella typhimurium TA102. Mutat. Res. 224:89-94,1989. 29. Wannamethee G, Ebrahim S, Shaper AG. Gamma-glutamyltransferase: determinants and association with mortality from ischemic heart disease and all causes. Am. J. Epidemiol. 142:699-708,1995. 30. Emdin M, Passino C, Mtehelassi C, Titta F, L'abbate A, Donate L, Pompella A, Paolicchi A. Prognostic value of serum gamma-glutamyl transferase activity after myocardial infarction. Eur. Heart J. 22:18027,2001. 31. Weinberg JM. The cell biology of ischemic renal injury. Kidney Int. 39:476-500,1991. 32. Bonventre JV. Mechanism of acute renal failure. Kidney Int. 43:1160-1178,1993. 33. Me Anulty JF, Huang XQ: The efficacy of antioxidants administered during low temperature storage of warm ischemic kidney tissue slices. Cryobiology 34:406-415,1997. 34. Eschwege P, Conti M, Paradis V, Pudliszewski M, Prieur E, Bendvald A, Bedossa P et al. Expression of aldehydic lipid peroxidation products in rat kidneys during warm ischemia. Transplant. Proc. 29: 2437-2438, 1997.

222

S. Dominici el al. / y-Glutamyltransferase-Dependent Prooxidant Reactions

35. Marathe GV, Nash B, Haschemeyer RH, Tate SS. Ultrastructural localization of gamma glutamyt transpeptidase in rat kidney and jejunum. FEBS Lett 107:436-440,1979. 36. Cutrin JC, Pompella A, Camandola S, Zingaro B, Sant S, Pol! G. Contribution of gamma-glutamyl transpeptidase to oxidative damage of iscnemic rat kidney. Kidney Int. 57:526-533,2000.

Thiol Metabolism and Redox Regulation of Cellular Functions A. Pompella et al. (Eds.) 1OS Press, 2002

223

Serum Gamma-Glutamyl Transpeptidase: a Prognostic Marker in Cardiovascular Diseases Michele EMDIN*, Claudio PASSING*, Alfonso POMPELLA*, and Aldo PAOLICCHI* Institute of Clinical Physiology - Via G. Moruzzi 1, Pisa, Italy, and # Dept. of Experimental Pathology, University of Pisa, Italy

1. Introduction Are arterial atherosclerosis and the haemodynamic consequences of a tight vessel stenosis sufficient to provoke the clinical ischemic syndrome, as always thought? Or something else occurs, triggering the most serious and fatal sequelae (unstable angina, acute myocardial infarction, stroke), towards irreversible organ damage and subsequent failure or sudden death [1]? The concept of instability of the plaque (vulnerable plaque), has been introduced not so long ago, to underline the behavioural difference among stenotic lesions with a similar lumen diameter reduction, in terms of provoking the clinical event. In other words, the "culprit" lesion might be not the "tightest" but the most "active" one. A variety of factors has been claimed, both clinical and humoral, with assessed prognostic significance, likely contributing to the evolution of the atherosclerotic lesion (by facilitating thrombosis, or plaque thickening and rupture); among others, inflammation has been pointed out as having a major role [2]. Plaques at risk for disruption tend to demonstrate outward vessel remodeling, to contain a large lipid core, thinned out fibrous cap, reduced collagen content, and increased inflammatory cell infiltration. To explain the beginning and the progression of the flogistic process, even a role for some "environmental" factors, such as viral or bacterial agents has been hypothesized [3], among many other possible triggers of proinflammatory cellular responses (e.g. cytokines, angiotensin II, hypertension, hyperglycemia, smoking, oxidative stress and modified LDL) [4]Low density lipoproteins (LDL) may per se induce inflammation whenever their peroxidation and oxidation/glycosylation of their protein moiety occur within the matrix of media layer, representing a stimulus for the endothelial cells to release chemokines, to promote monocyte adhesion and migration through the vessel wall into the plaque. The activated monocytes phagocytose the modified LDL, thus becoming "foam cells" which after death release lipids enlarging the plaque core. Proteolytic processes and oxygen related species destroy matrix molecules and damage muscle cells, which will not be able to repair the cap. The damage of the cap will start the thrombotic process. Oxidized LDL are actually lethal for various kinds of cells within the atheroma (myocytes, fibroblasts, macrophages, endothelial cells). Among the mechanisms able to induce in vitro LDL oxidation (mieloperoxidase, lipoxygenase, oxygen reactive species, Cu and Fe ions), the metabolism of thiols, and in particular of glutathione (GSH, gamma-glutamyl-cisteinyl-glycine, the main intracellular antioxidant agent) has been invoked as a possible trigger through the hydrolysis of its

224

M. Emdin et al. /Serum Gamma-Glutamyl Transpeptidase

y-glutamyl bond between glutamate and cysteine [5], through an extracellular reaction catalyzed by the gamma-glutamyltranspeptidase (y-GT). Through the action of membrane dipeptidases, this reaction provides cysteine and glycine to the intracellular milieu, as precursors for the GSH synthesis. Outside the cell, however, the cys-gly dipeptide is a powerful reductant for Fe3+, able to generate at the same time Fe2* and a free thiyl radical (Fig. 1). Furthermore, in the presence of GSH e Fe3+, y-GT is able to catalyze the LDL lipoprotein oxidation, at an enzyme activity similar to that currently found in human sera. While it is unlikely that this reaction takes place in plasma, due to the high levels of antioxidant agents, this is not the case of the plaque milieu, where free iron is present, and where Paolicchi et al. immuno-histochemically demonstrated the activity of y-GT, both in coronary and cerebrovascular human atheromas, colocalized with oxidized lipids and with inflammatory infiltration [5]. Serum GGT is partially absorbed onto LDL lipoproteins, which might carry y-GT activity inside the plaque.

GSH [pKa (SH) - 8.56]

gly-cys-SH [pKa (SH) = 6.4]

Fig. 1. Gamma-glutamyl transpeptldase (rGT) metabolism of glutathtone (GSH). Outside the cell, the CysGly dipeptide, is a powerful reductant for Fe3*. able to generate at the same time Fe2* and a free thiyl radical. Thereafter, oxygen reactive species, by the same reaction, contribute to a net prooxidant effect.

Another line of evidence points to iron metabolism as possibly linked with the evolution of the atherosclerotic process, as indicated by the association of increased iron body tissue reserve with an increased risk of myocardial infarction. Tissue iron content might be a relevant cofactor in influencing the predictive value of y-GT, thus suggesting adequate prospective studies [6]. 2. y-GT Role in Cardiovascular Diseases, an Emerging Perspective The serum levels of y-GT activity are currently considered as an index of hepato-biliary dysfunction and alcohol abuse [7]. Nevertheless, within its normal range, y-GT has many other, even stronger determinants than alcohol consumption. In sex-specific, multiple regression analyses [8,9], performed over large unselected populations on a total of 12511

M. Emdin el al. /Serum Gamma-Glutamyl Transpeptidase

225

men, and 12494 females, aged 12-62 years, screened in two different health survey programs, y-GT showed a strong positive association with body mass index, alcohol use, total serum cholesterol and a somewhat weaker positive association with serum triglycerides, high density lipoprotein cholesterol, heart rate, blood pressure, use of analgesics, time since last meal. Strong negative associations were found for coffee consumption, hour of the day at which the examination was performed and, in males, physical activity. In females, use of oral contraceptives and menopause were positively associated with serum y-GT, whereas pregnant females had lower values. Fewer than 3.85.5% of the males and 0.81.5% of the females had values exceeding 50 units/liter, as reported by these two distinct studies from Norway [8,9].

3. Prognostic Role of y-GT in Ischemic Heart Diseases An increasing number of population studies have evaluated the relation between serum y-GT activity and mortality, since the observation of Conigrave [10], indicating that y-GT does have a predictive value for mortality irrespective of hepatic disease or alcohol consumption. More recently, Jousilahti et al. analyzed the association of two widely recognized markers of alcohol consumption - carbohydrate-deficient transferrin (CDT) and y-GT - and self-reported alcohol consumption, with prevalent ischaemic heart disease in a random sample of 3666 Finnish men aged 25 to 74 years participating in a risk factor survey in 1997: finally, the CDT levels were inversely and y-GT levels positively correlated with CHD risk. In a composite risk assessment, men with normal CDT levels (z 20 U/L) and elevated y-GT levels (>80 U/L) had nearly 8-fold adjusted risk of ischaemic heart disease, as compared with the men with normal y-GT levels and elevated CDT levels [11]. Self-reported alcohol consumption had an inverse association with ischaemic heart disease risk, which disappeared after adjustment for the other risk factors. Thereafter, Wannamethee et al. [12] reported in a large unselected population of middle-aged men that y-GT has a negative prognostic value, as concerns both overall and cardiac mortality - namely mortality in patients with a previous history of ischemic heart disease - thus suggesting a linkage with underlying atherosclerotic coronary artery disease. In a prospective study of 7613 middle-aged British men followed for 11 years, y-GT levels were strongly associated with all-cause mortality, largely due to a significant increase in deaths from ischemic heart disease in the top quintile of the y-GT distribution. Serum y-GT levels were positively associated to preexisting ischemic heart disease, diabetes mellitus, antihypertensive medication, systolic and diastolic blood pressure, total and high density lipoprotein cholesterol, heart rate, and blood glucose, and negatively associated with physical activity and lung function. After adjustment for these variables, elevated y-GT (highest quintile, a24 unit/liter, vs. the rest) was still associated with a significant increase in mortality from all causes and from ischemic heart disease. The increased risk of ischemic heart disease mortality was more marked hi those with evidence of ischemic heart disease at screening, particularly in those with previous myocardial infarction. Another recent study was aimed at evaluating the long-term prognosis among 714 patients with a very small or unconfirmed acute myocardial infarction (AMI) aged <76 years: y-GT at multivariate analysis remained as independent risk indicator for death with age, previous myocardial infarction, smoking, and glucose [13]. The linkage with underlying atherosclerotic coronary artery disease has been recently demonstrated in a prospective study by our group, which evaluated a follow-up of over 6-years in 469 patients with ischemic syndrome and angiographically documented

226

M. Emdin et al. /Serum Gamma-Glutamyl Transpeptidase

coronary artery disease [14] (Fig. 2). After correction for other cardiovascular risk factors (such as age, smoking habit, serum cholesterol, left ventricular ejection fraction, body mass index, diabetes mellitus), or counfounding factors (such as serum ALAT and self-reported alcohol consumption), the prognostic value of serum Y-GT activity for cardiac death and non-fatal infarction was confirmed, in particular, in a subset of patients (corresponding to the 36% of the whole population) prone to plaque complications, characterized by the association of multivessel disease and an history of previous myocardial infarction irrespective of left ventricular function. The risk was increasing using different cut-off

100

82.1%

<40 U/L (327 patients)

80

2 00 CD _S

p=0.044

74.7%

60

>40 U/L (142 patients)

40H



w

0

12

24

36

48

60

72

Months Rg. 2. Event free survival, after 6 years of follow-up, according to serum yGT activity among 469 patients with coronary artery disease. Vertical lines represent confidence intervals.

values of 25 U/L or 40 U/L (however within the normal range), and the event excess was concentrated within the first three-year period (see Fig. 2). The prognostic significance of y-GT seems thus correlated not only with the extent of coronary artery disease, but also with the instability of the plaque. Importantly in fact, the prognostic significance of serum y-GT disappeared after revascularization procedures, in a subset of patients with a history of previous myocardial infarction and multivessel disease (Fig. 3). The observation of higher mean values of Y~GT in coronary artery disease patients as compared to the general population [8,14] points out its possible pathogenetic role in the first steps of atherosclerosis. 4. Y-GT and Ischemic Stroke Stroke mortality represents the third leading cause of death after coronary artery disease and all cancers. Various studies have reported a protective effect of light to moderate alcohol consumption on ischemic stroke risk [15].

M. Emdin et at. /Serum Gamma-Glutamyl Transpeptidase

221

There is substantial evidence that moderate drinking does not increase risk of ischemic stroke; however, studies remain divided on the question of a "protective" association. Furthermore, although the evidence is not unanimous, two major cohort studies have found that even moderate drinking may increase risk of hemorrhagic stroke. More recently, data from the Framingham study showed no significant association between alcohol intake and ischemic stroke overall, but showed a protective effect of alcohol among subjects aged 60 to 69 years [16].

12

24

36

48

89.4%

>40 U/L (41 patients)

85.0%

<40 U/L (142 patients)

60

72

Months Fig. 3. The subset of 262 patients with previous myocardlal Infarction was evaluated for occurrence of cardiac events. No difference was found according to a y-GT cutoff of 40 U/L.

A large Finnish study recently demonstrated that serum y-GT has an independent prognostic value for ischemic stroke in unselected populations in both genders, independently of self-reported alcohol drinking [17]. Authors focused their interpretation of data on the supposed link between the enzyme activity and alcohol consumption, which in their opinion might be underestimated by self-assessment through questionnaires. An alternative explanation of these findings has been given by us, underlying the possible pathogenetic role of y-GT in ischemic stroke in this cohort of subjects [18]. Another study evaluated the association between y-GT as a marker for alcohol consumption with fatal, non-fatal, haemorrhagic or ischemic stroke in three European cohort studies participating in the EUROSTROKE program. An increase of serum y-GT of one standard deviation (28.7 ILJ/ml) was associated with an age and sex adjusted 26% increase (95% CI: 5-53) in risk of stroke [19]. 5. Serum y-GT is a Prognostic Marker in Cardiovascular Diseases In conclusion, as concerns ischemic cerebral and heart disease, y-GT serum assay seems to have all the main features of a true prognostic marker:

228

M. Emdin et al. /Serum Gamma-Glutamyl Transpeptidase

1. optimal sensitivity-specificity of the diagnostic assay [6]; 2. epidemiological evidence of its presence in apparently healthy people before the occurrence of events [8,9]; 3. it really improves our ability to predict the event [11-14]; 4. finally, as concerns ischemic heart disease, though not essential, its prognostic impact can be affected by therapeutical intervention (revascularization), and this is associated with a decreased occurrence of cardiac events [14]. The recent new insights into the role of thiol methabolisn in atherosclerosis not only increase our understanding of the disease, but also have practical clinical applications in risk stratification and targeting of therapy for this clinical challenge of growing worldwide importance. Elevation of serum y-GT predicts outcomes of patients in unselected populations as well as of patients with ascertained ischemic heart disease, independently of myocardial damage, thus adding to prognostic information provided by other traditional risk factors. Prospectives studies will soon tell us how these findings relate with possible abnormalities in iron metabolism, or with markers of low-grade chronic inflammation, such as C-reactive protein, making thus feasible to dentify the most riskful combination for the vulnerable plaque, and to devise medical strategies for the stabilization of lesions, as an alternative approach to percutaneous or surgical procedures.

6. Conclusion From Dr. Jekill to Mr. Hyde [20], the physiological role of y-GT as a 'provider' of'fuel' for intracellular GSH resynthesis, and the interpretation of its increase in serum as the result of a compensatory overexpression in response to hepatobiliary dysfunction or alcohol toxic effect, have been recently revised in front of the evidence for a possible detrimental role, determined by the combination with multiple factors - some well recognized, as the extent of coronary atherosclerosis, others in need of further investigation, such as markers of inflammation. Acknowledgments The authors are indebted to Claudio Michelassi for precious help in the statistical evaluation of data. The graphical assistance of Luca Serasini is also gratefully acknowledged.

References 1. Shah PK. Pathophyskstogy of coronary thrombosis: role of plaque rupture and plaque erosion. Prog Cardiovasc Dis 2002; 44:357-68. 2. Libby P. Ridker PM, Maseri A. Inflammation and atheroscterosis.Circulation 2002;105:1135-43. 3. Shah PK. Link between infection and atherosclerosis. Who are the culprits: viruses, bacteria, both or neither? Circulation 2001; 103:5-7. 4. Ross R. Atherosclerosis: an inflammatory disease. N Engl J Med 1999; 340:115-126. 5. Paolicchi A, Minotti G, Tonarelli P, Tongiani R, De Cesare D, Mezzetti A, Dominici S, Comport M, Pompella A. Gamma-glutamyl transpeptidase-dependent iron reduction and low density lipoprotein

M. Emdin et al. /Serum Gamma-Glutamyl Transpeptidase

229

oxidation - a potential mechanism in atherosclerosis. J Invest Med. 1999^47:151-160. 6. Whitfield JB. Gamma-glutamyl transferase. Grit Rev Clin Lab Sci. 2001 ;38:263-355. 7. Tuomainen TP, Punnonen K, Nyyssonen K, Salonen JT. Association between body iron stores and the risk of acute myocardial infarction in men. Circulation 1998; 97:1461-6. 8. Arnesen E, Huseby NE, Brenn T, Try K. The Tromso Heart Study: distribution of, and determinants for, gamma-glutamyltransferase in a free-living population. Scand J Clin Lab Invest 1986; 46:63-70. 9. Nilssen O, Forde OH, Brenn T. The Tromso Study. Distribution and population determinants of gammaglutamyltransferase Am J Epidemiol 1990;132:318-26. 10. Conigrave KM, Saunders JB, Reznik RB, Whitfield JB. Prediction of alcohol-related harm by laboratory test results. Clin Chem. 1993;39:2266-70. 11. Jousilahti P, Vartiainen E, Alho H, Poikolainen K, Sillanaukee P. Opposite associations of carbohydratedeficient transferrin and gamma-glutamyltransferase with prevalent coronary heart disease. Arch Intern Med 2002; 162:817-21. 12. Wannamethee G, Ebrahim S, Shaper AG. Gamma-glutamyltransferase: determinants and association with mortality from ischaemic heart disease and all causes. Am J Epidemiol. 1995; 142:699-708. 13. Karlson BW, Wiklund O, Hallgren P, Sjolin M, Lindqvist J, Herlitz J. Ten-year mortality amongst patients with a very small or unconfirmed acute myocardial infarction in relation to clinical history, metabolic screening and signs of myocardial ischaemia. J Intern Med 2000; 247:449-56. 14. Emdin M, Passino C, Michelassi C, Titta F, L'Abbate A, Donate L, Pompella A, Paolicchi A. Prognostic value of serum gamma-glutamyl transferase activity in patients with ischaemic heart disease. Eur Heart J. 2001;22:1802-1807. 15. Camargo CA Jr. Case-control and cohort studies of moderate alcohol consumption and stroke. Clin Chim Acta 1996; 15;246:107-19. 16. Djousse L, Ellison RC, Beiser A, Scaramucci A, D'Agostino RB, Wolf PA. Alcohol consumption and risk of ischemic stroke: The Framingham Study. Stroke 2002; 33:907-12. 17. Jousilahti P, Rastenyte D, Tuomilehto J. Serum gamma-glutamyl transferase, self-reported alcohol drinking, and the risk of stroke. Stroke 2000:3:1851-5. 18. Emdin M, Passino C, Donato L, Paolicchi A, Pompella A. Serum gamma-glutamyltransferase as a risk factor of ischaemic might be independent of alcohol consumption. Stroke, 2002, 33:1163-4. 19. Bots ML, Salonen JT, Elwood PC, Nikitin Y, Freire de Concalves A, Inzitari D, Sivenius J, Trichopoulou A, Tuomilehto J, Koudstaal PJ, Grobbee DE. Gamma-glutamyltransferase and risk of stroke: the EUROSTROKE project. J Epidemiol Community Health 2002;56 Suppl 1 :i25-9. 20. Stevenson RL. Dr. Jekyll and Mr. Hyde, 1886.

230

Thiol Metabolism and Redox Regulation of Cellular Functions A. Pompella etal. (Eds.) IOS Press, 2002

Lipoic Acid: a Multifunctional Antioxidant Aalt BAST and Guido R.M.M. HAENEN University Maastricht, Faculty of Medicine, Department of Pharmacology and Toxicology, P.O. Box 616, 6200 MD Maastricht, The Netherlands. 1.Introduction Lipoic acid (LA) has gained attention in the late 1940's as a growth factor for microorganisms. In 1951 Reed and co-workers [1] isolated 30 mg of this factor form 100 kg of liver and later on the molecular structure of LA (chemical name: 1,2-dithiolane-3-pentanoic acid) was elucidated. The compound has a chiral center at the C3 carbon atom (Fig. 1). Naturally occurring LA has the R configuration.

OH

Lipoic acid (LA) (0»\

O

OH SH

Dihydrolipoic acid (DHLA)

Rg. 1. The molecular structure of llpote acid and dlhydrolipoic acid

LA is present in most pro- and eukaryotic cells. In our diet it can be found in several products such as meat, liver and heart. LA is considered not to be a vitamin because of possible biosynthesis in man. The physiological function of LA as a component in the pyruvate dehydrogenase complex is now textbook knowledge. In the enzyme complexes LA is linked by an amide bound to the y-amino group of a lysine residue of the protein, giving the so-called lipoyl group. In this way the dithiolane ring is attached to the enzyme by a relatively long and flexible chain, which makes it possible for this group to react with the catalytic center of other enzymes in the enzyme complexes. 2. Antioxidant action The applications of LA as antioxidant has rekindled scientific interest in LA, in particular the need to understand the influence of LA on patho-physiological processes. Compared to open chain disulfides, such as glutathione disulfide (GSSG), LA is a better antioxidant. For example,

,4. Bast and G.R.M.M. Haenen / Lipoic Acid: a Multifunctional Antioxidant

231

the ability of LA to protect ai-antiproteinase against inactivation by hypochlorous acid is comparable to that of reduced glutathione (GSH) and other thiols. GSSG is not able to protect against this hypochlorous acid-induced damage. The explanation for the difference in scavenging activity between both disulfides can be found in the dihedral angle of the C-S-S-C moiety, cp. In open chain disulfides, as GSSG, cp approximates 90°. At this angle the energetically unfavorable interaction between the sulfur lone pairs is kept at a minimum. The disulfide group in LA is in a dithiolane, a strained, five membered ring. In the dithiolane cp is 35°. This results in an energetically unfavorable interaction of the orbital of the lone pz electron pairs of the sulfur atoms. The relatively high energy content of the disulfide group in LA is responsible for the relatively high reactivity of LA with oxidizing species. The dihedral angle, cp, of a disulfide in a 6 membered ring is 60°, which will reduce the reactivity of such a disulfide compared to that in LA [2]. The relatively high reactivity of the five membered ring is probably the reason why nature has selected a dithiolane-containing compound as cofactor in several enzyme complexes. The antioxidant activity of the reduced form of LA (Fig. 1), dihydrolipoic acid (DHLA), is superior to that of LA. Therefore, much interest has been focussed on the reduction of LA. LA can be reduced to DHLA by several enzymes. In mitochondria lipoamide dehydrogenase catalyses the oxidation of a dihydrolipoyl group to a lipoyl group. In this reaction NAD+ is converted into NADH. Lipoamide dehydrogenase can also catalyze the reduction of LA at the expense of NADH [3]. In the cytosol, GSSG-reductase can reduce LA. In this reaction NADPH is consumed. Both enzymes have an opposite stereospecificity: Mammalian GSSG-reductase preferentially reduces the S-enantiomer whereas lipoamide dehydrogenase preferentially reduces the R-enantiomer [3].

3. A Racemic Mixture As a drug, LA is administered as a racemic mixture. Remarkably, the administration of racemic LA to experimental animals yielded more DHLA than administration of either the pure Renantiomer or the S-enantiomer. This can be explained by the fast, non-enzymatic dithioldisulfide exchange between both enantiomers. Lipoamide dehydrogenase preferentially reduces R-LA. In a subsequent reaction, the formed R-DHLA may non-enzymatically reduce S-LA. A similar synergistic effect of S-LA on the reduction of R-LA by GSSG-reductase can be expected. The synergistic effect that both enantiomers can have on the reduction of the other enantiomer can explain the higher DHLA levels found after administration of the racemic mixture compared to that after administration of only one of the enantiomers [4]. Of course, additional clinical studies are needed to substantiate that the racemic mixture has a superior therapeutic effect. To our knowledge, LA would than be the first drug of which - despite stereospecific enzymatic activation - the use of a racemic mixture is preferred over the use of one of the enantiomers [5]. It is, however, known that the R-enantiomer has a better bioavailability than the S-enantiomer. 4. Two Thiol Groups The direct antioxidant activity of DHLA resides in its thiol groups. The chemical reactivity of the thiol groups in DHLA is, however, less than that of GSH. In the nucleophilic substitution

232

A. Bast and G.R.M.M. Haenen /Lipoic Acid: a Multifunctional Antioxidant

reactions with l-chloro-2,4-dinitrobenzene and chloroacetamide, the second order rate constant for GSH is 2 to 3 times higher than that for DHLA (Fig. 2). Moreover, the cellular concentration of GSH exceeds the maximal concentration of DHLA that can be reached after therapeutic application by far. Additionally, it should be noted that enzymes such as GSH-Stransferases and GSH-peroxidases further enhance the reactivity of GSH. This indicates that a therapeutic effect of DHLA cannot be achieved by the chemical reactivity of the thiol groups itself.

cr*

SH

SH

CI'+H* ^SH

NO2

Fig. 2. The nucteophllic substitution reaction of glutathtone (GSH) and dlhydrollpoic acid (DHLA). The second order rate constant for the reaction of the GSH with 1-chloro-2,4dinrtrobenzene (at 37 °C and pH 7.4) is 2.00 • 1O5 M/min and for DHLA this is 1.00 10* M/rnin.

There are, however, examples of reactions where the activity of DHLA is superior to that of GSH. A well-studied example is the interaction of thiols with ebselen [6]. Ebselen is a low molecular weight organo-selenium compound that mimics the activity of the selenium containing GSH-peroxidase. In its peroxidase activity, ebselen is first reduced to the corresponding selenol (Fig. 3). In the actual peroxidase reaction, the selenol reacts with an hydroperoxide and ebselen is formed again. Two thiols are needed to reduce ebselen to the selenol. In the reaction with GSH, first a selenenyl sulfide is formed (Fig. 3). Subsequently, this intermediate reacts with a second GSH giving the selenol and GSSG. In the conversion of ebselen to the selenol, the reaction of ebselen with the first thiol is fast. The reaction of the intermediary selenenylsulfide with the second GSH appears to be rate limiting. Using DHLA in stead of GSH as reductor drastically increases the rate of selenol formation. The intermediary selenenylsulfide of the reaction between ebselen and DHLA contains an intramolecular thiol group (Fig. 3). This intramolecular thiol group reacts instantaneously with the sulfur of the selenium-sulfur moiety, giving ebselen and LA. An intramolecular reaction of the ebselen-DHLA intermediate is much faster than an intramolecular reaction of the ebselenGSH adduct with another GSH. Based on these findings it can be concluded that a unique activity of DHLA can be found in reductions where two thiol groups are needed, provided that the reaction with the second thiol is rate limiting. Only in this type of reaction the reactivity of DHLA is superior to that of GSH. In these reactions the presence of two vicinal thiol groups in one molecule that

A. Bast and G.R.M.M. Haenen /Lipoic Acid: a Multifunctional Antioxidant

233

can form an intramolecular disulfide explains the potent activity of DHLA. A comparison can be made to dithiothreitol, a reducing dithiol that is used extensively in biochemistry. DHLA can be considered as the endogenous dithiothreitol.

ROH + H20

ROOM

DHLA ROH + H2O

\ ^SeDHLA selenenyl sulfide

ROOM

B. Fig. 3. The molecular mechanism of the peroxldase activity of ebselen with (glutathione) GSH (A) or with dlhydrollpolc acid (DHLA) as cofactor (B). With GSH the rate limiting step is the formation of the selenol from the selenyl sulfide, whereas with DHLA the formation of the selenol is very fast due to the presence of an intramolecular thiol group in the DHLA-selenenyl sulfide.

Also in the interaction with other antioxidants in the antioxidant screen (vide infra), the unique reducing capacity of DHLA comes into play. In the regeneration of ascorbate from dehydroascorbate, DHLA is superior to monothiols like GSH. It is also involved in the DHLA mediated reduction of GSSG to GSH. In the fast dithiol-disulfide exchange between both enantiomers, the potent reducing capacity of DHLA, as well as the high reactivity of the disulfide groups in LA are important.

234

A. Bast and G.R.M.M. Haenen /Lipoic Acid: a Multifunctional Antioxidant

at normal glucose levels glucose + ATP

hexokinase . , ^^ > glucose-6-phosphate + ADP

glucose-6-phosphate + NADP++ H+ *• 6-phosphoglucono-8-lactone + NADPH glucose-6-phosphate dehydrogenase

GSSG

1 GSH A.

at high glucose levels aldose reductase NADPH + H + glucose > +

NAD+ + sorbitol

sorbitol + NADP+

> fructose + NADH + H+ sorbitol dehydrogenase

lipoic acid

a dihydrolipoic acid

i

GSSG

GSH

B.

Fig. 4. At physiological glucose levels (Fig. 4a), glucose produces NADPH that can be used for the reduction the glutathlone disulflde (GSSG) to GSH. High glucose levels shift the production of NADPH to NADH (Fig. 4b). This will hamper the reduction of GSSG. However, in the presence of lipoic acid (and lipoamide dehydrogenase) NADH can be used to give dihydrolipoic add, which enables the reduction of GSSG.

A. Bast and G.R.M.M. Haenen / Lipoic Acid: a Multifunctional Antioxidant

235

5. Preventive and Curative Another example of the extraordinary reactivity of DHLA is the repair of oxidatively damaged methionine residues. Several oxidants, e.g. hypochlorous acid, preferentially oxidize methionine residues. The enzyme peptide bound methionine sulfoxide reductase (PMRS) is able to repair this oxidative damage. It was found that DHLA can enhance the PMRS mediated repair by supplying the necessary reducing equivalents. Both LA and DHLA can protect aiantiproteinase against hypochlorous acid-induced inactivation by scavenging hypochlorous acid. The extra antioxidant activity of DHLA is that it promotes the repair of oxidized aiantiproteinase by enhancing PMSR activity [7]. Interestingly, the latter antioxidant action is not preventive but curative.

6. Therapeutic Actions A remarkably diverse range of actions can be ascribed to LA. Besides the ability of the dithiol as well as the disulphide form to act as antioxidants, the dithiol group can also be used for the regeneration of vitamin E, vitamin C and glutathione [8]. Both the sulphur atoms and the carboxylic moiety are involved in complexing metals [2]. This explains its use in intoxications. Both the activation and the nuclear translocation and action of NF-KB are inhibited by the redox couple dihydrolipoic acid and lipoic acid [9] LA is successfully used in the treatment of type II diabetes induced polyneuropathy. Several mechanisms that explain this therapeutic mode of action have been suggested [10,11]. DHLA has been suggested to substitute for HS-CoA in various enzymatic reactions and to decrease acetyl-CoA. LA increases the uptake of glucose. The increase in glucose uptake has been explained by an enhanced GLUT-4 glucose transporter function. Since the GLUT-4 increase is not accompanied by an increase in its mRNA, it is presumed that the degradation of GLUT4 is decreased by LA. In hyperglycemia, glucose-utilizing enzymes become saturated and glucose is irreversibly reduced to sorbitol by aldose reductase consuming NADPH. Sorbitol is then oxidized to fructose forming NADH. This so-called polyol pathway, therefore shifts reducing equivalents from NADPH to NADH (Fig. 4). LA is reduced by lipoamide reductase which is dependent of NADH. At high glucose levels the NADPH depletion hampers the glutathione reductase catalysed reduction of GSSG. In this case, DHLA can reduce GSSG. The reduction of GSSG through LA - via DHLA - uses NADH for this reaction. Thus, lipoic acid also alleviates the NADH surplus in diabetes. DHLA prevents the formation of advanced glycation end products (AGE's). This is probably explained by the sequestration of reactive

o

o

Fig. 5. The molecular structure of the major metabolite of lipoic acid, 3-ketollpolc acid.

236

A. Bast and G.R.M.M. Haenen /Lipoic Acid: a Multifunctional Antioxidant

aldehydes by the thiols groups in DHL A. Moreover, Bierhaus et al. [12] have shown that the incubation of endothelial cells with AGE's gave intracellular oxidative stress that could be prevented by LA. In addition, glucose led to an increase in NF-KB, which also can be prevented by LA. This exemplifies that LA can prevent glucose mediated damage via various mechanisms. Many studies have been performed with the racemate of lipoic acid (R,S-lipoic acid). The bioavailability and the action on glucose uptake are better for the R-enantiomer [13]. We found that a major metabolite of R-lipoic acid is 3-ketolipoic acid (Fig. 5). Its molecular structure reveals that this compound might be regarded as a bioactive metabolite. The contribution of this metabolite to the therapeutic effect of LA has not been elucidated thus far.

7. Conclusions LA is characterized by ring strain in the dithiolane group. The close proximity of the two thiol groups renders unique properties to the dithiol in reactions where two thiol groups are needed. The clinically established action of LA in prevention of complications of type II diabetes invites for further biochemical research.

References [1] Reed L.J., De Busk B.G., Gunsalus I.C., Schnakenberg G.H.F.

Crystalline a-lipote acid: A catalytic agent

asociated with pyruvate dehydrogenase. Science 114, 93 (1951). [2] Biewenga G.Ph., Haenen G.R.M.M., Bast A. An overview of lipoate chemistry. In: Lipoic acid in health and disease. Eds.J. Fuchs, L. Packer, G. Zimmer. Macel Dekker Inc. New York Pp. 1-32 (1997). [2] Biewenga G. Ph., Haenen G.R.M.M., Bast A.

The pharmacology of the antioxidant lipoic acid.

Gen.

Pharmacol. 29, 315-331 (1997). [3] Biewenga G.Ph., Dorstijn M.A., Verhagen J.V., Haenen G.R.M.M., Bast A. The reduction of lipoic acid by lipoamide dehydrogenase. Biochem. Pharmacol. 51, 233-238 (1996). [4] Maitra I., Serbinova E., Tritschler H., Packer L Alpha-lipoic acid prevents buthionine sutfoximine-induced cataract formation in newborn rats. Free Rad. Biol. Med. 18, 823-829 (1995). [5] Biewenga G.Ph., Haenen G.R.M.M., Groen B.H., Biewenga J.E., van Grondelle R., Bast A. Combined nonenzymatic and enzymatic reduction favours bioactivation of racemic lipoic acid. An advantage of a racemic drug? Chirality 9, 362-366 (1997). [6] Haenen G.R.M.M., de Rooij B.M., Vermeuten N.P.E.. Bast A. Mechanism of the reaction of ebseten with endogenous thiols: Dihydrolipoate is a better cofactor than glutathione in the peroxkJase activity of ebseten. Mol. Pharmacol. 37, 412-422 (1990). [7] Biewenga G.Ph., Veening-Griffioen D.H., Nicastia A.J., Haenen G.R.M.M., Bast A. A new antioxidant property of dihydrolipoic acid: Repair of oxkJatively damaged alpha-1 antiprotease. Drug Res. 48,144-148 (1998). [8] Bast A., Haenen G.R.M.M. Interplay between glutathione and lipoic acid in the protection against microsomal lipkJ peroxidation. Biochim. Biophys. Acta 963, 558-561 (1988). [9] Packer L. Alpha-lipoic acid: A metabolic antioxidant which regulates NF-kappa B signal transduction and protects against oxidative injury. Drug Metab. Rev. 30, 245-275 (1998). [10] Biewenga G. Ph., Haenen G.R.M.M., Bast A. The role of lipoic acid in the treatment of diabetic polyneuropathy. Drug Metab. Rev. 29, 1025-1054 (1998).

A. Bast and G.R.M.M. Haenen /Lipoic Acid: a Multifunctional Antioxidant

[11] Biewenga G.Ph., Haenen G.R.M.M.,

237

Bast A. The pharmacology of the antioxidant lipoic acid. Gen.

Pharmacol. 29, 315-331 (1997). [12] Bierhaus A, Chevion S., Chevion M., Hofmann M., Quehenberger P., Illmer T., Luther T., Berentshtein E., Tritschler H., Muller M., Wahl P., Ziegler R., Nawroth P.P. Advanced glycation end product-induced activation of NF-KB is suppressed by a-lipoic acid in cultured endothelial cells. Diabetes 46,1481-1490 (1997). [13] Hermann R., Niebch G. Human pharmacokinetics of a-lipoic acid.. In: Lipoic acid in health and disease. Eds.J. Fuchs, L. Packer, G. Zimmer. Macel Dekker Inc. New York Pp. 337-360 (1997).

238

Thiol Metabolism and Redox Regulation of Cellular Functions A. Pompella eial. (Eds.) IOS Press, 2002

Is Glutathione an Important Neuroprotective Effector Molecule against Amyloid Beta Toxicity ? Vicki S.BARBER and Helen R.GRIFFITHS PSRI, Aston University, Aston Triangle, Birmingham B4 JET, UK

1. Introduction Increasing age is the most reliable and robust risk factor for susceptibility to neurodegenerative disease (Bains & Shaw, 1997). Lovell et al., (1995) showed support for the concept that the brain in Alzheimer's disease (AD) is under increased oxidative stress (OS) demonstrating lipid peroxidation changes in areas where degenerative changes occur. Further evidence suggesting that the pathogenesis of AD is as a result of oxidative damage includes elevated levels of iron in AD brains and a colocalised reduction in antioxidant status (Good et al., 1996). Whilst the exact mechanisms underlying oxidative stress remain unclear, it has been proposed that the peptidergic fragment of amyloid beta (A0) that accumulates in AD may exert its toxicity through peroxide generation (Huang et al., 1999). Glutathione is arguably the most important AOX and free radical scavenger present in cells (Valencia et al, 2001), where glutathione-associated metabolism is a major mechanism for cellular protection against agents that generate OS. Glutathione participates in detoxification at several different levels, and may scavenge reactive oxygen species (ROS), reduce peroxides, or be conjugated with electrophilic compounds. Thus, glutathione provides the cell with multiple defences not only against ROS but also against their toxic products. Most importantly, many of the glutathione-dependent proteins are inducible and therefore represent a means whereby cells can adapt to OS (Hayes & McLellan, 1999). The GSH redox status is critical for various biological events that include transcriptional activation of specific genes, modulation of redox-regulated signal transduction, regulation of cell proliferation, apoptosis, and inflammation (Rahman & MacNee, 2000). In addition, it has been shown previously that GSH levels decrease following addition of cytotoxic agents, and at the time of onset of apoptosis (van den Dobbelsteen et al., 1996; Froissard & Duval, 1994). The intracellular synthesis of GSH is mainly regulated by gamma glutamyl cysteinyl synthetase (y-GCS) (Richman & Meister, 1975). Differences or alterations in the levels of protein can occur by a number of different mechanisms including alterations in the level of gene transcription and alterations in the stability or translatability of the resulting RNA. The expression of y-GCS is sensitive to OS, where the existence of the OS-response element, AP-1, on the y-GCSh (y-GCS heavy subunit) promoter has been clarified (Mulcahy90%) minimises the accumulation of disulfides and provides a reducing environment within the cell. If there is a shift in the GSH/GSSG redox buffer, a variety of cellular signalling processes are influenced, such as activation and phosphorylation of stress kinases (JNK, p38, PI-3K) via sensitive cysteine-rich domains; activation of sphingomyelinase-ceramide pathway, and

V.S. Barber and H.R. Griffiths / I s Glutathione a Neuroprotective Effector Molecule?

239

activation of the transcription factors AP-1 and NF-icB, eventually leading to increased gene transcription (Rahman & MacNee, 2000). Froissard et al., (1997) have reported that there is strong evidence to support the existence of a close relationship between the level of intracellular glutathione and cell survival in PC 12 cells, where mitochondrial GSH is critical for the maintenance of mitochondrial function and cellular viability (Seyfried et al., (1999). However, whether a drop in GSH levels precede intracellular ROS production during apoptosis or vice versa is not certain. Hence any role of GSH depletion alone in triggering apoptosis is not clear (Hall, 1999). Evidence is accumulating for an important role for glutathione in detoxification processes in the brain. Depletion of glutathione in newborn rats using buthionine sulfoximine (BSO) leads to mitochondrial damage in the brain (Jain et al , 1991). Furthermore, application of beta amyloid peptide, glutamate receptor agonists and BSO cause GSH depletion in cultured neurones and lead to induction of apoptosis. HaOz has previously been suggested to mediate A|3 cytotoxicity based on antioxidant inhibition of toxicity and demonstration of lipid peroxidation in treated cells (Behl et al., 1994b). The toxic effects of several other ROS generating compounds has been shown to be abrogated by neurotrophic factors such as BDNF, NGF, GDNF and bFGF (Cheng and Mattson, 1995; Chao and Lee). It has been postulated that this may be due to upregulation of the concentration of glutathione and/or the activities of antioxidant enzymes. Therefore we have investigated the hypothesis that Af3 neurotoxicity is mediated by reactive oxygen species, where trophic factor cytoprotection against oxidative stress is achieved through regulation of glutathione levels at the gene level. Herein, we demonstrate that AJ3 toxicity is associated with alterations in cellular redox status, and whilst NGF affords protection against toxicity, this is independent of GCS expression. 2. Methods Maintenance of PC 12 Cell Line PC 12 cells were routinely cultured in 75cm2 flasks in a water jacketed humidified 5% CO2 incubator with maintenance media. (RPMI 1640 media with Glutamax I; 10% (v/v) HS, 5% (v/v) FBS, penicillin (0.5U/ml) and streptomycin (0.5mg/ml). Every 2-3 days, spent PC 12 cell media was removed and replaced with fresh maintenance media, which had been pre-warmed to 37°C. Cells were passaged once a week, as described in section 2.2.1.3, when the cells were at a density of approximately 5 x 106 per ml. For experimental purposes, cells were seeded at 2 x 105 cells per ml and allowed to rest for 4 hours prior to treatment for times and doses indicated. Determination of Cell Viability using 3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium (MTT) Assay Cells (2 x 105 cells per ml, lOOul per well) were seeded into 96 well flat bottomed microtitre plates. Two hours prior to completion of the experiment, MTT solution (25ul of 5mg/ml in 0.01M PBS) was added to all wells including blanks. Plates were then incubated for a further 2 hours at 37°C, 5% CO2. Lysis buffer (lOO^il of 20% w/v SDS, in DMF (50%), dH2O(50%), pH 4.7 adjusted with 2.5% of 80% glacial acetic acid) was then added to each well and the plates incubated for a further 16 hours at 37°C, 5%CO2. The absorbance of the each well was then read at 570nm in a 96 well plate reader.

240

V.S. Barber and H.R. Griffiths /Is Glutathione a Neuroprotective Effector Molecule?

Dichlorofluorescein Diacetate (DCFDA) Staining Cells (2 x 105 per ml, 2ml) were incubated with test agents in 35mm well plates. 30 minutes before the end of incubation, DCFDA solution (20ul of 7.5mM diluted stock in PBS) was added to the plates. After exactly 30 minutes cells were immediately harvested with a Gilson pipette and a rubber policeman and transferred to a flow cytometry tube. Samples were then immediately run through a Coulter EPICS flow cytometer with the intensity of light scatter and the FL1 fluorescence emitted between 505 and 535nm following excitation at 488nm (argon laser) was recorded for individual nucleoids. The median X was calculated for 10,000 nucleoid events. Glutathione (GSH) Assay Glutathione (reduced and oxidised) levels were determined using modified microtitre plate method (Punchard et a/., 1994). Actual glutathione levels were calculated by equating the changes in absorbance over time alongside standards of known GSH concentrations included on each plate. Glutathione (GSH), reduced standards were freshly prepared for each experiment from a stock (lOOmM in sterile distilled water). Cells (2x10 ) were then harvested and PBS (1ml) added to wash the pellet. Tubes were recentrifuged at 6600 x g for 1.5 minutes and the PBS was then removed, making sure to leave the pellet dry. SSA (3.33ul of 1% made up in distilled water) was then added to precipitate the protein and the tubes were immediately centrifuged at 13000 x g for 1.5 minutes. Stock buffer (96.6ul of 125mM sodium phosphate, 6.3mM disodium EDTA, pH 7.5 autoclaved) was then added to each tube. Standards and samples (25 uJ) were then ah*quoted into a flat bottomed plates (96 well) in triplicate, the remaining aliquot was immediately frozen at -70°C for later use in the GSSG assay. Daily buffer (150ul of 0.3mg NADPH/ml stock buffer) and DTNB solution (50u,l of 6mM DTNB in daily buffer) was then added to all wells. The plate was then loaded onto the carriage of a 96 well plate reader and glutathione reductase solution (25ul of 20U/ml) was then added as quickly as possible to initiate the reaction. The OD was then recorded every minute for 5 minutes at 410nm. Glutathione (GSSG) Assay Oxidised glutathione (GSSG) standards were freshly prepared for each experiment from an oxidised glutathione stock (1.5uM in sterile distilled water). The remaining 25ul aliquot from the GSH assay was employed for this assay. To the frozen aliquot and standards (25uJ), 2-vinylpyridine (2-VP), triethanolamine (O.Sul) were added. Samples and standards were then vortexed and centrifuged for 30 seconds at 13000 x g. Aliquots of standards and samples (lOul) were then plated out into a microtitre plate (96 well plate). Daily buffer (75uJ of 0.3mg NADPH/ml stock buffer) and DTNB solution (25uJ of 6mM DTNB in daily buffer) were added to the wells. The plate was then loaded onto the carriage of a 96 well plate reader and glutathione reductase solution (12.5ul of 20U/ml) was then added as quickly as possible to initiate the reaction. The OD was then recorded every minute for 5 minutes at 4 lOnm. Estimation of Protein Concentration Protein concentrations were assessed according to a modified version of Sigma Procedure No. TPRO-562, based on the method of Smith et al, (1985). Protein standards of six concentrations were prepared from the protein standard solution (bovine serum albumin,

V.S. Barber and H.R. Griffiths / I s Glutathione a Neuroprotective Effector Molecule?

241

lmg/ml) diluted in coating buffer (15mM sodium carbonate, 35mM sodium hydrogen carbonate, pH 9.2). Each standard (lOul) was plated out in triplicate into a 96 well flat bottomed microtitre plate. Unknown samples (lOul) were also plated out in triplicate. Bicinchoninic acid solution (200ul) containing copper (II) sulphate pentahydrate 4% solution (50:1) was then added to all wells (samples and standards), the plate was then incubated at 37°C for 30 minutes. The absorbance of the plate was then measured at 570nm in a 96 well plate reader. Unknowns were calculated from a standard curve prepared using the software package GraphPad PRISM. mRNA Extraction Cells (2xl05) were harvested into a RT grade centrifuge tube (1.5ml). Cells were centrifuged for 1.5minutes at 13000 x g in an Eppendorf Centrifuge 5415D. Superaatants were removed and PBS (1ml) added, cells were centrifuged for another 1.5minutes at 13000 x g and the PBS removed. The extraction was carried out using the Dynal mRNA direct kit protocol. Pellets were resuspended into 200 Dl Dynal lysis/binding buffer (lOOmM Tris-HCl, pH 7.5, 500mM LiCl, lOmM EDTA, pHS.O, 1% LiDS and 5mM DTT) and aspirated twenty times to fully resuspend each pellet with fresh tips being used for each tube. The solution was then aspirated five times using 1ml sterile syringes and sterile 21 gauge needles to shear the DNA. This was then repeated using 1ml sterile syringes and sterile 25 gauge needles. Tubes were then centrifuged for 1 minute at 13000 x g in an Eppendorf Centrifuge 5415D. Resuspended Dynal Oligo (dT)2s beads (30ul) were washed in Dynal lysis/binding buffer twice before being added to each tube of lysed cells. The tubes were then mixed on a Dynal sample mixer for 5 minutes at 22-25°C. Tubes were then magnetised on the Dynal MPC®-E magnet (Magnetic particle concentrator for microtubes of Eppendorf type (1.5ml)) and the colourless solution removed. Dynal wash buffer A (200ul of lOmM Tris-HCl, pH 7.5, 0.15M LiCl, ImM EDTA, 0.1% LiDS) was then added to the pellet and the cells resuspended in it. The tubes were then remagnetised the colourless solution removed and Dynal wash buffer A added (200ul) and the cells resuspended. The process was then repeated using Dynal wash buffer B (200ui of lOmM Tris-HCl, pH 7.5, 0.15M LiCl, ImM EDTA). After two washes with wash buffer B the cell pellet was resuspended in DEPC treated water (30pl of 0.1%). This was then transferred to a fresh tube (1.5ml), and master mix was added (39ul of lOmM DTT (7.5ul of lOOmM), ImM dNTPs (7.5ul dNTP mix), 25U RNAsin (l.Sul), 1U (3ul) RQ1 RNase-free DNase, 0.1% DEPC treated water (4.2ul) and Expand buffer (15ul)). Each tube was aspirated twice and incubated at 37°C for 60 minutes. DNase was heat inactivated at 70°C for 10 minutes in a preheated dry heating block. Samples were then centrifuged at 13000 x g for 1 minute in an Eppendorf centrifuge 5415D. Reverse Transcription of mRNA Two RT grade PCR tubes were labelled for each sample, one positive and one negative. RNAsin 15U (lu,l) was added to the negative tubes and 30U (2ul) to the positive tubes, mRNA extraction product (46ul) was added to the positive tubes and mRNA extraction product (23 ul) was added to the negative tubes. Expand RT SOU, (lul) was then added to the positive tubes. For reverse transcription mRNA was incubated at 37°C for 1 hour in Expand RT buffer containing lOmM DTT, ImM dNTPs, 25U RNAsin, 1U RQ1 RNasefree DNase. DNase was heat inactivated at 70°C for 10 minutes. The samples were then incubated at 42°C for 1 hour before being stored in the short term at 0 - 4°C, and at -20°C in the long term.

242

V.S. Barber and H.R. Griffiths /Is Glutathione a Neuroprotective Effector Molecule?

Glyceraldehyde 3 Phosphate Dehydrogenase (GAPDH) and GCS Pofymerase Reaction (PCR)

Chain

The primers (SOnmol of each) as stated in table 2.4 were synthesised by Gibco/BRL, Life Technologies, Paisley, Scotland. Each primer was diluted to 100 pmol/ul in sterile 1 x TE buffer (lOmM Tris, ImM EOT A, pH 8.0), and then further diluted tenfold in 1 x TE buffer and stored as lOuJ aliquots at -20°C. Primer Sequences Primer

Forward primer sequence (61 - 3')

Reverse primer sequence (V-V)

Reference

GAPDH

AGA ACA TCA TCC CTG CCT C

GCC AAA TTC GTT GTC ATA CC

Hal etal.. 1998

GCS

CCT TCT GGC ACA GCA CGT TG

TAA GAC GGC ATC TCG CTC CT

El Mouatassiuin et al , 2000

PCR was performed in PCR buffer with sterile water, 200mM dNTPs and lOpmol of each primer. All reactions were covered in a drop of mineral oil and hot start conditions were used, full details of each PCR as stated in table 2.5. Reactions were initiated with 2.5U (O.SuJ) Taq DNA Polymerase. All liquid handling was conducted using filter tips. Optimal PCR Conditions

PRIMER

CONDITIONS

CYCLES

GAPDH

98°C 3 minutes + Taq 60°C 2 minutes 72°C 2 minutes

1 cycle

26 cycles 94°C 30 seconds 60°C 30 seconds 72°C 30 seconds 94°C 30 seconds 60°C 30 seconds 72°C 4 minutes

GCS

94°C 94°C + Taq 56°C 72°C

1 minute 45 seconds

1 cycle

1 cycle

1 minute 1 minute

94°C 45 seconds 56°C 1 minute 72°C 1 minute 72°C 10 minutes

31 cycles

1 cycle

Agarose gel electrophoresis of PCR products Gel loading solution (2.5ul of 0.05% w/v bromophenol blue, 40% w/v sucrose, 0.1M EDTA, pH 8.0) was added to the PCR reaction products (10D1). Samples were loaded onto

V.S. Barber and H.R. Griffiths /Is Glutathione a Neuroprotective Effector Molecule?

243

1.5% agarose gels made up in 1 x TBE (89mM Tris, 89mM boric acid, 2.5mM EOT A, pH 8.0, 100ml Gel). Gels were electrophoresed at 80V for 2 hours in 350ml 1 x TBE. Post staining was achieved with ethidium bromide (lOul of a lOmg/ml solution) dissolved into 100ml of distilled water for 15 minutes in the dark at 22 - 25°C. Gels were visualised on an U.V. transilluminator table at 312nm. Expression levels were quantitated using the software package Phoretix ID Advanced, Version 4.01 (Non Linear Dynamic Limited) and normalised to GAPDH. Gels were all loaded with a size ladder, usually a 123bp DNA ladder prepared according to manufacturer's instructions.

3. Results It has been proposed that AP may exert its toxicity through peroxide generation. In order to confirm whether generation of ROS occurs in PC 12 cells and whether these levels are modulated by exogenous agents, the effects of AJ3 on intracellular peroxide levels were investigated in PC 12 in the presence and absence of NGF. PC 12 cells were incubated with NGF (5 and lOng/ml) for 24 hours before addition of AP (25 uM) and the fluorochrome DCFDA. Cells were harvested and immediately analysed through a flow cytometer. Exposure of PC12 cells to Ap 25-35 (lOuM) caused an increase in intracellular peroxide which was evident after 4 hours. Production of intracellular peroxide was inhibited by either desferioxamine or NGF (lOng/ml), where NGF had no effect on basal cellular peroxide in control cells (see Figure 1). Pre-treatment with 5ng/ml NGF prior to Ap elicited a decrease in the median X value for DCF fluorescence to 0.300, indicative of a decrease in intracellular peroxide concentration (data not shown).

X

0.8-

TJ

0.4-

0.0

(O

m CL + x UL Q

o> o

X u_ O

O)

o "-

Treatment Fig. 1. PC12 cells (2x10 ) were incubated for 4 hours at 37°C, 5%COz before the addition of NGF (10ng/ml) or DFX (100uM) for 24 hours at 37°C, 5%CO2 Ap (25uM) was then added for 4 hours at 370C, 5%CO2. For the final 30 minutes of incubation, DCFDA (75uM) was added. Cells were then harvested and analysed by flow cytometry. Figure shows the average median X ± SEM.

244

V.S. Barber and H.R. Griffiths /Is Glutathione a Neuroprotective Effector Molecule?

In order to establish whether intracellular ROS production was associated with a reduction in cell viability, MTT reductive capacity was investigated after exposure to AfJ or ^Oj. In the absence of NGF, A3 (lOuM) caused a 25% reduction in MTT reductive capacity after 24 hours treatment, and this effects was not enhanced by higher peptide concentrations. Much higher concentrations of hydrogen peroxide were required to elicit the same loss of viability; after 24hours treatment with 150uM HjC^, MTT reductive capacity was reduced by 23% (Figure 2a and 2b).

10

0

25

Amyloid p (|iM) Fig. 2a. The effects of amyloid p (25-35) toxicity on PC12 cells. PC12 cells (2x105) were incubated with aged Ap (10, and 25uM) for 24 hours at 37°C, 5% CO2. Cells were then subject to the MTT assay. Graph shows the mean absorbance of formazan produced ± SEM for at least 3 replicates. ** represents P<0.01 (Dunnetf s test).

1.00n

^—

-**-

_l_

** —r—

-rtr

_ 0.75-

**

*

•o 0.500

o

0.25-

n nn-

0

50 100 Hydrogen Peroxide

150

Fig. 2b. The effect of hydrogen peroxide on PC12 cell viability using the MTT assay. PC12 cells (2 x 10s) were incubated for 4 hours at 37°C, 5% CCh before the addition of HsA? (50. 100. and 150jiM) for 24 hours at 37°C, 5% CO*. Cells were then subject to the MTT assay. Data shown are the mean absorbance of formazan produced ± SEM of 6 replicates for each treatment. ** represents P values of <0.01 (Dunnett's test).

Previously, others have shown that NGF is an effective cytoprotectant against oxidative toxicity arising from glutamate or dopamine. However, this has not been shown for hydrogen peroxide or amyloid beta. Figure 3 demonstrates that NGF affords protection to PC 12 cells against peroxide toxicity, when cells have been previously exposed to 5 or

V.S. Barber and H.R. Griffiths /Is Glutathione a Neuroprotective Effector Molecule?

245

10ng/ml NGF prior to oxidant challenge. Higher doses of NGF (50ng/ml) exerted a cytotoxic effect (data not shown).

CZU Ong/ml NGF CUD 5ng/ml NGF •11 Ong/ml NGF

0

160

Hydrogen Peroxide Treatment Fig. 3. The effects of nerve growth factor on PC12 cell viability in the presence or absence of HzOz. PC12 cells (2 x 10s) were incubated with NGF (5 and 10ng/ml) for 24 hours at 37°C, 5% COs. H2O2 (150>M) was then added for a further 24 hours at 37°C, 5% CO2. Cells were then subject to the standard MTT assay. Graph shows the mean absorbance of formazan produced ± SEM for at least 14 replicates.

Glutathione participates in detoxification at several different levels, and can scavenge ROS and reduce peroxides. To determine whether intracellular protection is compromised by oxidative challenge, and to determine the kinetics of glutathione metabolism during 24hours of oxidative stress treatment elicited by either hydrogen peroxide or AJ3, total cellular glutathione was measured after 1, 4 and 24 hours. Figure 4 illustrates the rapid loss of intracellular GSH following addition of hydrogen peroxide to PC 12 cells in culture, where levels reduced by 50% within one hour. In contrast, A(3 did not affect GSH levels until four hours after treatment. Nevertheless, A0 treatments did cause a twofold increase in total intracellular glutathione in the surviving cell population at 24 hours (p<0.05), where peroxide induced a highly significant fourfold increase in GSH (P<0.001). Again after 24 hours peroxide treatment (160uM), there was a ten fold increase oxidised glutathione, but the concentration of GSSG only rose from 0.02% to 0.05% of total glutathione (data not shown). In order to determine whether part of the cytoprotective effects of NGF could be attributed to up-regulation of glutathione, the effects of NGF in the presence and absence of oxidant challenge on glutathione levels and on expression of GCS mRNA were examined. Figure 6 shows NGF to prevent the increase in GSH levels observed post- exposure to A(3, where 1 Ong/ml NGF significantly reduced intracellular glutathione concentration after 4 and 24 hours respectively. Analysis of GCS mRNA expression in PC 12 cells exposed to oxidative stress in the presence or absence of NGF is shown in figure 7. In the absence of peroxide challenge, NGF elicited an increase in GCS mRNA expression after two hours in PC 12 cells, whereas mRNA levels were essentially unchanged in the absence of NGF. However, when PC 12 cells were subject to an oxidative challenge for 4 hours after 24hours pre-incubation with vehicle or NGF, cells showed reduced expression of GCS mRNA irrespective of the presence or absence of NGF. Furthermore, 24hours after oxidative challenge in the presence in NGF, GCS mRNA expression was restored to that seen prior to oxidative

246

V.S. Barber and H.R. Griffiths / Is Glutathione a Neuroprotective Effector

Molecule?

challenge. In contrast, PC 12 cells subject to oxidative challenge in the absence of NGF showed up-regulation of GCS expression 24 hours post-oxidative challenge, where image analysis using phoretix software demonstrated a twelve-fold increase in expression. CZZZ30 hours 1 hour 14 hour 124 hour

PBS

peroxide

Treatment Fig. 4. Kinetics of hydrogen peroxide effect* on glutathione levels in PC12 cells. PC12 cells (2x105) were incubated with t-feOz (150 uM) for 0-24 hours at 37°C, SttCCfe. Figure shows the mean glutathione value ± SEM of 3 replicates. *** represents P<0 006 (Bonferroni's test) 200-1

cm OUM APES iouMAp

n

0.0

A 4.0

24.0

Hours Fig. 5. The effects of amyloid p peptide on glutathione levels in PC12 cells. PC 12 cells (2x10s) were incubated with Apaws (10|iM) for 0-24 hours at 37°C, 5% CCfe. Figure shows the mean glutathione value ± SEM of at least 4 replicates. * represents P<0.05 (Dunnetfs test)

CZl Ong/rri NGF + OuM Ap BB 10ng/rrt NGF + OuM Ap E3 Ong/rrt NGF + 10uM Ap CH 10ng/rrt NGF + 10uM Ap

4. 24. Hours of Treatment Fig. 6a. The effects of nerve growth factor on amyloid p pepttde induced changes to glutathione levels in PC12 cells. PC12 cells (2x105) were incubated with NGF (10ng/ml) for 24 hours at 37°C. 5% CO2 before incubation with AP25-35 (10nM) for 0-24 hours at 37°C. 5% CO2. Figure shows the mean glutathione value ± SEM of at least 3 replicates. * represents P<0 05 (Unpaired West).

V.S. Barber and H.R. Griffiths / Is Glutathione a Neuroprotective Effector Molecule ?

247

CZDOng/rrtNGF ••Sng/mlNGF •HOng/mlNGF

0

50

100

150

Hydrogen Peroxide Treatment

Fig. 6b. The effects of nerve growth factor on the action of hydrogen peroxide on glutathione levels in PC12 cells. PC12 cells (2x105) were incubated with NGF (5 and 10ng/ml) for 24 hours at 37°C, 5%CO2 before addition of H2O2 (50, 100, ISO^M) at 37°C, 5% CO for 24 hours2. Figure shows the mean glutathione value ± SEM of 3 replicates.

4. Discussion The levels of intracellular peroxide can be quantitated using the nonfluorescent fluorochrome 2',7'-dichlorofluorescin diacetate (DCFDA) which is freely permeable to cells. It is de-esterified within cells by endogenous esterases to 2',7'-dichlorofluorescein (DCF). This becomes trapped within cells and accumulated DCF is able to interact with peroxides and hydroperoxides in the cell to form the fluorescent compound 2',7'dichlorofluorescein (Behl et al., 1994a; Zhu et al., 1994). There is an important caveat to the successful use of DCF as a peroxide sensor, and that it is also a suitable substrate for direct oxidation by xanthine oxidase, where inhibition by allopurinol can be used to determine any role of the latter. DCF staining has previously been shown to increase with ROS production (Satoh et al., 1997). Addition of BbO2 increases intracellular peroxide in PC12 cells (data not shown); this is in agreement with the work of other groups (Jang & Surh, 2001). Herein, we have shown that Af3 increases intracellular peroxides. The inclusion of the iron chelator DFX suggests this process is iron dependent, and suggests that AJ3 cannot exert its toxicity without the availability of iron. These observations support the work of Monji et al., (2002) who have demonstrated that Ap (25-35)-associated free radical generation is iron dependent and strongly influenced by the aggregational state of the peptides Whilst NGF pre-treatment in unchallenged cells did not cause any significant lowering of intracellular peroxide levels in PC 12 cells, it was effective in preventing the elevation in cytoplasmic peroxide observed following amyloid beta challenge. Previously trophic factors have been shown to protect cells against ROS toxicity in cultured neurones and PC 12 cells, however these papers did not measure intracellular ROS directly (Chang and Mattson; 1995; Gong et al, 1999). In addition, these earlier studies have investigated oxidative stressors implicated in the pathogenesis of Parkinson's disease. Amyloid beta is implicated in cholinergic neurone loss in AD, and in our system, we have demonstrated the generation of ROS precedes cell death. Again, in our model, the toxic effects of amyloid beta were prevented by NGF.

V. S. Barber ami H. R. Griffiths / Is Glutathione a Neuroprotective Effector Molecule ?

248

+ NGF TREATMENT 1 2 3 4 5 6 7 8 9

No NGF TREATMENT 1 2 3 4 5 6 7 8

369bp 246bp 123bp Treatment

Lane

Treatment

Lane

1

123bp ladder

1

1 23 bp ladder

2

Blank

2

Blank

3

OhrNGF

3

1 hour PBS

4

1 hr NGF

4

2 hr PBS

5

2hr NGF

5

4 hr PBS

6

4hrNGF

6

24 hr PBS

7

24 hr NGF

7

24hr PBS + 4 hr hbCb

8

24hr NGF+ 4hr HjC^

8

24 hr PBS + 24 hr KfcOj

9

24hr NGF+24hr HjOz

9

Fig. 7. 1.5% Agarose gels of PCR products. PC12 cells (2x105) were incubated with NGF (10ng/ml) or PBS as control for 0-24 hours at 37°C, 5%CO2 H2O2 (150jiM) was then added for 4 and 24 hours at 37°C, 5%CO2 and cells were harvested and mRNA extracted. Folowing reverse transcription specific PCRs were underatkaen for GAPDH and GCS.

It has been postulated that the cytoprotective effect of trophic factors may be dependent on the upregulation of glutathione and/or enzymes associated with ROS defence. Examination of the effects of ROS on cellular glutathione confirmed the observations of others, that GSH is rapidly consumed, but then recovers to offer enhanced protection against further insults. OS appears to be a key pleiotropic modulator of gene transcription acting through the generation of bioactive mediators (Morel & Barouki, 1999). The number of transcription factors whose activities are modulated by ROS is substantial, and include AP-1, p53, NF-xJB, HSF, Sp-1 and GABP (Kehrer, 2000). In this respect, ROS appear to be actual modulators of gene transcription, independently of the degradation of biological macromolecules (Morel & Barouki, 1999). Therefore if ROS can induce an adaptive response, for example though induction of GCS, offering greater protection of the cell exposed to subsequent OS. In this paper, we report that NGF (lOng/ml) did not cause any significant changes in GSH levels of unchallenged cells; this is in agreement with Kamata et a/., (19%) who showed NGF to protect cells from OS independently of GSH. However under H2O2 and A(3 challenge, increased levels of GSH were recorded versus control after 24 hours, with NGF (lOng/ml) preventing the elevation induced by oxidative stress. The data recorded with 5ng/ml NGF is consistent with the data previously published by Pan & Perez-Polo

V.S. Barber and H.R. Griffiths / I s Glutathione a Neuroprotective Effector Molecule?

249

(1993) and Sampath et al, (1994) who have both shown NGF to cause increases in GSH levels. GSSG levels were also assessed where HbC^ to caused increased GSSG concentration after challenge with 160uM peroxide. Nevertheless, it is evident from this data that much of the GSH is lost particularly at early time points, either through mixed disulphide formation with proteins or through efflux. The drop in GSH may therefore reflect an inhibition of synthesis, a deficit in the GSH salvage pathway, or an increased rate of efflux (Chandra et al, 2000). In both scenarios, efficient removal of ROS and formation of GSSG is dependent on expression and activity of GPX, which in itself is redox sensitive, and the activity of the enzyme under oxidative stress. Specifically, the effect of NGF merits further investigation, since infusion of GDNF into the brain has been shown to increase the activity of GPx (Chao and Lee, 1999) Yonezawa et al., (1996) have shown that deprivation of one of the precursors of glutathione (cysteine) induces glutathione depletion and death in cultures of cells, but that rescue by free radical scavenger and a diffusible glial factor was not associated with a restoration of normal glutathione content. These authors thus suggest that the protecting agents act distal to glutathione. This is in agreement with Kamata et al., (1996) who showed when cellular GSH was depleted by treatment with BSO, NGF still protected cells. These data largely support the hypothesis that NGF exerts its protection independently of GSH. NGF (lOng/ml) does not to cause any significant increases in glutathione levels, however, does protect against intracellular ROS and also cell death. To further clarify whether NGF exerts its protection via glutathione up regulation, studies on GCS expression were undertaken. Cell cross-talk between glia and neurones appears critical to maintainence of neuronal viabililty. Glia conditioned medium has been shown to contain an unidentified diffusible factor which enhances resistance to oxidative stress by increasing transcription of GCS (Iwata-Ichikawa, 1999). In order to determine whether NGF can elicit a similar response in PC 12 cells, we have evaluated GCS expression in PC 12 cells, in relation to glyceraldehydes 3-phosphate expression (GAPDH). The glycolytic tetramer GAPDH/G3PDH is a multifunctional enzyme involved in cellular metabolism. It is an important glycolytic enzyme that catalyses the oxidative phosphorylation of glyceraldehyde-3-phosphate to 1,3-diphosphoglycerate, and it is the most widely used internal control gene (Suzuki et al., 2000). Measurement of GAPDH transcript abundance was originally selected as a normalizer for other expressions because it encodes for a protein with a housekeeping function. As it is a glycolytic intermediate it is therefore expected to be present in all cells and exhibit minimal modulation. However GAPDH gene expression has been reported to increase during other non-glycolytic activities such as programmed neuronal cell death (Ishitani et al., 1996). During investigations into selective GCS gene expression, GAPDH expression remained constant. GCS expression was more variable in NGF treated cells compared to non-NGF treated cells. The only marked differences in non NGF treated cells were after 4 hours of peroxide challenge where expression was virtually nothing and after 24 hours of peroxide challenge where there was the greatest expression. In NGF treated cells GCS levels initially increased up to 2 hours after the NGF treatment then decreased. After 4 hours of peroxide challenge GCS levels were also greatly reduced showing the least expression, expression then increased after 24 hours of peroxide challenge. Tubulin is frequently used as a housekeeping gene for neuronal studies, however, transcription of this species is itself affected by the onset of apoptosis, rendering it a poor control gene. One of the most variable factors appears to lie in the variation in GSH levels within resting cells. Indeed, GSH levels have also been shown to decrease in culture imposing the need for consistent experimental conditions (Reiners Jr et al., 2000), including passage frequency and seeding density.

250

V.S. Barber and H.R. Griffiths/ Is Glutathione a Neuroprotective Effector Molecule?

There is a poor correlation between mRNA and protein levels, as the presence of RNA in a cell is not necessarily indicative of peptide synthesis. Structurally authentic mRNAs may be poorly handled by the translational machinery in a cell. Even the efficient translation of a functional message may not result in biologically active peptide if post translational processes are absent, or if the peptide is rapidly degraded. The analysis of GCS gene expression must therefore extend to the description of the protein end-points of the expression pathway, included in this must be a consideration of protein function (Carter & Murphy, 1999). References Bains JS and Shaw CA. Neurodegenerative disorders in humans: the role of glutathione in oxidative stressmediated neuronal death, Brain Research Reviews, Volume 25, Issue 3, December 1997, Pages 335-358. Chandra J, Samali A and Orrenius S. Hydrogen peroxide mediates amyloid beta protein toxicity, Cell, Volume 77, Issue6, 1994, Pages 817-827. Behl C, Davis JB, Lesley R and Schubert D. Triggering and modulation of apoptosis by oxidative stress. Free Radical Biology and Medicine, Volume 29, Issues 3-4, Pages 323-333. Cheng B and Mattson MP. PDGPs protect hippocampal neurons against energy deprivation and oxidative injury: evidence for induction of antioxidant pathways, Journal of Neuroscience: the Official Journal of the Society for Neuroscience, Volume 15, Issue 11, Pages 7095-7104. Chao CC and Lee EHY. Neuroprotective mechanism of glial cell line-derived neurotrophic factor on dopamine neurons: role of antioxidation, Neuropharmacotogy, Volume 38, Issue 6, 15 June 1999, Pages 913-916 Froissard, P; Duval, D. Cytotoxic effects of glutamic acid on PC12 cells, Neurochemistry International, Volume 24, Issue 5, May 1994, Pages 485-493. Froissard P, Monrocq H and Duval D. Role of glutathione metabolism in the glutamate-induced programmed cell death of neuronaHike PC12 cells, European Journal of Pharmacology, Volume 326, Issue 1, 1997, Pages 93-99. Gong L, Wyatt RJ, Baker I and Masserano JM. Brain-derived and glial cell line-derived neurotrophic factors protect a catecholaminergic cell line from dopamine-induced cell death, Neuroscience Letters, Volume 263, Issue 2-3,1999. Pages 153-156. Good PF, Werner P, Hsu A, danow CW and Perl DP. Evidence of neuronal oxidative damage in Alzheimer's disease, American Journal of Pathology, Volume 149, Issue 1, Pages 21-28. Hall AG. The role of glutathione in the regulation of apoptosis, European Journal of Clinical Investigation, Volume 29, Issue 3, Pages 238-245. Hayes JD and McLellan LI. Glutathione and glutathione-dependent enzymes represent a co-ordinately regulated defence against oxidative stress, Free Radical Research, Volume 31, Issue 4, Pages 273-300. Huang X, Atwood CS, Hartshorn MA, Multhaup G, Goldstein LE, Scarpa RC, Cuajungco MP, Gray DN, Lim J et al. The A beta peptide of Alzheimer's disease directly produces hydrogen peroxide through metal ion reduction, Biochemistry, Volume 38, Issue 24,1999, Pages 7609-7616. Iwata-lchikawa E, Kondo Y, Miyazaki I, Asanuma M and Ogawa N. Glial cells protect neurons against oxidative stress via transcriptional up-regulation of the glutathione synthesis, Journal of Neurochemistry, Volume 72, Issue 6, Pages 2334-2344. Jain A, Martensson J, Stole E, Auld PA and Meister A. Glutathione deficiency leads to mitochondrial damage in brain. P.N.A.S. USA, Volume 88, Issue 5,1991, Pages 1913-1917. Jang J and Sum Y. Protective effects of resveratrol on hydrogen peroxide-induced apoptosis in rat pheochromocytoma (PC12) cells, Mutation Research, Volume 496, Issue 1-2, 2001, Pages 181-190. Kamata H, Tanaka C, Yagisawa H and Hirata H. Nerve growth factor and forskolin prevent H2O2-induced apoptosis in PC12 cells by glutathione independent mechanism, Neuroscience Letters, Volume 212, Issue 3. 1996, Pages 179-182. Monji A, Utsumi H, Tadashi Ueda T, Imoto T, Yoshida I, Hashioka S, Tashiro K and Tashiro N. Amytoid-pprotein (Ap~) (25~35)-associated free radical generation is strongly influenced by the aggregational state of the peptides, Life Sciences, Volume 70, Issue 7, Pages 833-841. Morel Y and Barouki R. Repression of gene expression by oxidative stress, The Biochemical Journal, Volume 342 Part 3, 1999, Pages 481-496. Mulcahy RT, Wartman MA, Bailey HH and Gipp JJ. Constitutive and beta-naphthoflavone-induced expression of the human gamma-glutamylcysteine synthetase heavy subunit gene is regulated by a distal antioxidant response element/TRE sequence, Journal of Biological Chemistry, Volume 272, Issue 11, 1997. Pages 7445-7454. Pan Z and Perez-Polo R. Role of nerve growth factor in oxidant homeostasis: glutathione metabolism, Journal of Neurochemistry. Volume 61, Issue 5, Pages 1713-1721. Rahman I and MacNee W. Oxidative stress and regulation of glutathione in lung inflammation. The European Respiratory Journal, Volume 16, Issue 3, September 2000. Pages 534-554. Richman PG and Meister A. Regulation of gamma-glutamyl-cysteine synthetase by nonallosteric feedback inhibition by glutathione, Journal of Biological Chemistry, Volume 250, Issue 4, 1975, Pages 1422-1426

V.S. Barber and H.R. Griffiths / Is Glutathione a Neuroprotective Effector Molecule':

Sampath D, Jackson GR, Werrbach-Perez K and Perez-Polo JR. Effects of nerve growth factor on glutathione peroxidase and catalase in PC12 cells, Journal of Neurochemistry, Volume 62, Issue 6, Pages 2476-2479. Satoh T, Enokido Y, Aoshima H, Uchiyama Y and Hatanaka H. Changes in mitochondria! membrane potential during oxidative stress-induced apoptosis in PC12 cells, Journal of Neuroscience Research, Volume 50, Issue 3, 1997, Pages 413-420. Seyfried J, Soldner F, Schulz JB, Klockgether T, Kova, KA and Wullner, U. Differential effects of L-buthionine sulfoximine and ethacrynic acid on glutathione levels and mitochondria! function in PC12 cells, Neuroscience Letters, Volume 264, Issue 1-3,1999, Pages 1-4. Valencia E, Marin A and Hardy G. Glutathione—nutritional and pharmacologic viewpoints: part I, Nutrition, Volume 17, Pages 428-429. van den Oobbelsteen DJ, Nobel CS, Schlegel J, Cotgreave IA, Orrenius S and Slater AF. Rapid and specific efflux of reduced glutathione during apoptosis induced by anti-Fas/APO-1 antibody, Journal of Biological Chemistry, Volume 271, Issue 26, 1996, Pages 15420-15427. Zhu H, Bannenberg GL, Moldeus P and Shertzer HG. Oxidation pathways for the intracellular probe 2', 7'dichlorofluorescein, Archives of Toxicology, Volume 68, Issue 9, 1994, Pages 582-587.

252

Thiol Metabolism and Redox Regulation of Cellular Functions A. Pompella et al. (Eds.) IOS Press, 2002

Antioxidants in Cancer Therapy: is there a Rationale to Recommend Antioxidants during Cancer Therapy? Hans Konrad BIESALSKI and Jiirgen FRANK Dept. of Biological Chemistry and Nutrition, University of Hohenheim Fruwirthstrasse 12-D 70593 Stuttgart, Germany - email: biesal(qjiini-hohenheim. de

1. Introduction Antioxidants are regarded to be safe and healthy in general and to protect cells and tissues from all kinds of environmental and endogenous attacks. This preventive and protective image is generally accepted in the medical and "layman" community. Quite recently, the question has arisen whether concurrent administration of oral antioxidants is contraindicated during cancer treatment, since antioxidants might reduce oxidizing free radicals created by radiotherapy, photodynamic therapy, and some forms of chemotherapy, and thereby decrease the effectiveness of these treatments. Often, cancer therapy produces ROS which attacks healthy cells and tissues consequently leads to further damage and unintentional side effects. These adverse effects may be decreased by oral antioxidants, given before or simultaneously with tumor treatment. Therefore, there is a conflicting view of antioxidant use in cancer therapy. A recent review summarises the mechanisms of anti cancer agents and their relationship to oxidative stress [1]. The positive image of antioxidants does obviously not create the question whether this protection may be also available for cancer cells which subsequently would or could reduce the therapeutic efficacy. In the following paper therapeutic intervention of cancer based on ROS generation will be critically discussed in some but not all cancer therapies with respect to adjuvant antioxidant therapy.

2. Antioxidants and Cancer The frequent observation that suggested precancerous lesions such as adenomatous polyps, metaplastic or dysplastic areas, are low hi antioxidants, especially endogenous like vitamin E and carotenoids, leads to the assumption, that a low tissue levels of these antioxidants as a result of low intake contributes or are the main cause of cancer [2,3]. Indeed, epidemiological studies show that a low intake of fruit and vegetable protects from cancer (WCRF). However, this does not mean that the preventive effect of antioxidants is also present in established cancer. The initiation of cancer via oxidative stress (low antioxidants in relationship to high prooxidants) is believed to be an important step. In contrast, data exist, showing that agents that generate free radicals can selectively kill cancer cells, an effect which is blocked by antioxidants resulting in an acceleration cancer growth in vitro and in vivo [4-7]. As a consequence, antioxidants should not always perceived from their sunny side as major protecting agents.

H. K. Biesalski and J. Frank /Antioxidants in Cancer Therapy

253

Besides the exogenous antioxidants (vit. C., E, carotenoids etc.) there are three major types of primary intracellular antioxidant enzymes in mammalian cells: superoxide dismutase (SOD), catalase (CAT) and peroxidases, of which glutathione peroxidase is the most prominent. Cancer cells are frequently described to be low in MnSOD (mainly localized in mitochondria), CuZu SOD (in cytoplasma and nucleus) and CAT [8;9]. Indeed, increasing SOD activity in cancer cells via administration in liposomes or cDNAtransfection leads to tumor suppression in vitro. Additional treatment with l,3bis- (2chloroethyil)l-nitrosourea (BCNU) increases the cytotoxicity of BCNU [10]. This might be due to the BCNU sensitizing effect of tumor cells (glioma) against H2O2. Indeed, the sensitivity of tumor cells to HiOi induced oxidative stress can be greatly enhanced by impairing the antioxidant repair mechanisms in glioma but not normal brain cells with compounds like the buthionine sulfoxamine (BSO) [11]. Results from the above mentioned studies rise the question whether the frequently recommended combination of antioxidants with cancer therapy, to protect healthy tissues, might also protect tumor cells against free radicals, derived from either chemo- or radiotherapy or other treatments. Labriola and Livingstone (1999) [12] discuss six factors which may predict interactions between chemotherapeutic agents and antioxidant compounds Factors that may be useful in predicting antioxidant-chemotherapeutic drug interactions: - extent to which effectiveness of chemotherapeutic agent depends on generation of reactive oxygen species; - nature of reactive oxygen species generated by chemotherapeutic agent; - dosage and concentration of reactive oxygen species; - nature of the antioxidant; - concentration of the antioxidant; - temporal relationship between use of antioxidant and administration of chemotherapy. Damage of tumor cell DNA via free radicals plays an important role in cancer chemotherapy. However, many cancer cells are resistant to radical based DNA damage produced by either chemo- or radiotherapy [13; 14] due to a subpopulation of tumor cells that are hypoxic [15; 16]. Under low oxygen concentrations (relative to comparable normal cells) the cytotoxicity of a DNA radical lesion is decreased because a couple of DNA radical lesions require prior reaction with molecular oxygen followed by fragmentation [17]. With respect to this an important factor is the specific metabolic (energetic) condition of the tumor itself. Differences in oxygenation within the tumor tissue may be used to improve the therapeutic approach. These differences ranging from pO2 0 to pOa >80 mm Hg may act like an intrinsic ischemia/reperfusion system. The ischemia/reperfusion condition, e.g. areas with changing oxygen pressure, is frequently present in tumor tissues. This condition is characterized by transient ischemia, resulting in ATP hydrolysis, hypoxanthine accumulation and transformation of xanthine dehydrogenase to xanthine oxidase (XO), followed by reperfusion with high oxygen pressure. The generation of superoxide anion radicals (Oa'") and hydrogen peroxide (H^Oa) during the oxidation of hypoxanthine by XO is well established and is believed to play an important role in tissue damage caused by ischemia/reperfusion. Both, the presence of XO and elevated hypoxanthine concentrations has been described in tumor tissues [18]. Due to the impaired energy metabolism in tumors, the accumulation of ADP as well as the degradation of nucleic acids following cell death, promote the conditions for the production of significant amounts of superoxide anion radicals during the XO reaction. The superoxide dismutation product, hydrogen peroxide, is a well-established apoptotic agent [19] and may therefore increase the effect of XO. Furthermore, the reaction of reactive oxygen species with low

254

"• K- Biesalski and J. Frank /Antioxidants in Cancer Therapy

molecular weight iron complexes would easily result in the onset of lipid peroxidation [20], and indeed the release of such iron complexes is likely to occur during disaggregation of dead cells. Thus, a series of biochemical changes set into motion within tumor tissue by either free radical formation or by boosting the ischemia/reperfusion condition would concur to favour the occurrence of oxidative cell injury and apoptosis.

3. Radiation Radiation (x-ray, y-ray, PDT) is used to treat solid tumors either via photoexcitation (low energy) or ionizing radiation (high energy). The effectiveness of radiation is the result of DNA damage of tumor cells with subsequent apoptosis. This is mediated via ROSformation (hydroxyl radicals, singlett oxygen etc.). It is estimated that about 65% of the damage is a result of hydroxyl radical formation [21]. Depending on the energy of radiation (high or low) different sensitizers (radiation sensitizers or chromophores) are used to optimize the effect of the treatment. Radiation sensitizers are used to increase the sensitivity of hypoxic tumor cells against radiotherapy. 2-Nitro-imidazoles have a strong one electron reduction potential and are able to increase the radiation induced DNA damage especially in cases of hypoxic cells (KOV 34). In hypoxic cells the 2-Nitroimidazoles act like radiosensitizers by mimicking the damaging effect of molecular oxygen [22]. Under aerobic conditions, 2-nitro-imidazole is first reduced to the nitro radical anion. Together with superoxide anions the cytotoxic peroxynitite (ONOO~) anion can be formed. As depicted in figure 1A the formation of ONOO" results in nitrosylation of proteins preferably in endothelial cells. The ONOO* can in fact react with aromatic amino acids, probably through the formation of the nitronium cation NO2+ and of the NO2 radical. This primarily results in the addition of nitrate groups to the ortho position of tyrosines, a process referred to as 'protein nitration' [23]. Figure IB shows a more or less complete nitrosylation of endothelial cells following PDT-induced ROS-formation. But also under hypoxic conditions these nitro-derivatives express a high toxicity. Brezden and co workers demonstrated [24] that l-methyl-2-nitroimidazole depletes glutathione and protein thiols under hypoxic and aerobic conditions. Even nucleosomal DNA fragmentation did not occur, characteristics of apoptotic cell death (zeiosis and chromatin condensation) were present. Whether signs of apoptosis were a result of intracellular depletion of sulfhydryls and subsequently enhanced oxidative stress has to be further elucidated. Tirapazamine exhibits selective toxicity towards hypoxic cells via formation of a radical species which results in DNA damage. Under aerobic conditions, activated tirapazamine is rapidly reoxidized by molecular oxygen with concomitant formation of superoxide and ROS. The cytotoxicity of tirapazamine is due to its ability to generate free radicals on the dexoribase moiety and then to act as a surrogate for molecular oxygen by donating an oxygen atom to the deoxyribosyl radical. At low levels of tirapazamine intracellular antioxidants such as SOD, GSH and CAT mitigate the effect cytotoxicity of ROS formed [25]. Further compounds such as heterocyclic nitro compounds, quinones, methotrexate, paraquate, actinomycin, adriamycin, phenothiazine, acridinium salts as well as compounds of metals fall into typical electron transfer categories, e.g. cisplatin cytotoxicity increases with decreases in GSH-levels [26] pointing on the dependence of the antioxidant defence and effectiveness of treatment.

H.K. Biesalski and J. Frank/Antioxidants in Cancer Therapy

255

R-SH (thiol compound) peroxymtnte anion
OH

tyrosine

visuaJisation of "protein nitration" by immuno3-nitTOtyrosine histochemistry

B Fig. 1. (A) Reactions of reactive nitrogen species, nitroxide radical and peroxynitrite anion, leading to the formation of adducts on protein tyrosine residues ("protein nitration), which can be visualised by immunohisochemistry. (B) Simultaneous visualisation of nitrosylated proteins (left: green fluorescence) and von Willebrand factor (right: red fluorescence) after induction of oxidative stress. Nuclei were stained with DAPI (blue fluorescence).

4. Photodynamic Therapy Photodynamic therapy (PDT) entails the combined use of light and photosensitizers. This kind of treatment results in the formation of ROS and their effects on cellular structures such as mitochondria [27], lipids, proteins and DNA [28;29]. Similar to x-ray and combined chemotherapeutics to increase the radiosensitivity, PDT combines a photosensitizer with low energy radiation to treat tumors in the presence of molecular oxygen. As a consequence of direct transfer of energy from a photoexcited photosensitizer to molecular oxygen, a veriety of cytotoxic reactive oxygen species are created, the principle one being singlet oxygen. A prerequisite for these reactions however, is the presence of oxygen. Due to differences in oxygenation, a high rate of PDT-induced oxygen consumption regions distant from the capillary may become depleted of oxygen

256

H.K. Biesalski and J. Frank / Antioxidants in Cancer Therapy

during PDT treatment. If however singlet oxygen quencher such as. carotenoids accumulate in tumor tissues they may further reduce the efficacy of PDT treatment. PDT-mediated oxidative stress can function as a molecular switch for selective and temporal expression of heterologous genes (TNF-a, p53) in tumor cells under the control of a heat shock protein promotor [30]. This expression depends on the presence of singlet oxygen mediated oxidative stress. Nuclear transkription factor-kappa B (NFicB) [31;32] transkription factor AP-1 [32], transcription factor AP-2 [33], certain protein cinases [34;35], and phosphatases [36], have also been implicated in the regulation of ROSmediated gene expression. Depending on the pattern of gene expression induced, cellular responses may range from inflammation, degradation and irnmunosuppression to triggering of apoptosis [32;37]. Whether the frequentiy described modulating effects of antioxidants on cellular stress response including heat shock proteins, hemoxygenase, and their transcription factors (e.g. NFicB, AP-1 and -2, HSF-1) is also of importance in tumor cells and may subsequently interfere with the effectiveness of PDT is unknown, but should be taken in mind. In addition it was recently described that PDT-treatment of activated lymphocytes results in an adaptive antioxidative response (increase of GSHPx, increase of CAT activity and diminution of GSH/GSSH ratio) an effect which may contribute to a defence of the tumor cell against PDT induced oxidative stress.

5. Metal-containing Compounds Bleomycin acts via formation of ROS, a ferric peroxide as well as via hydroxyl radicals and induces apoptosis. Hug and coworkers (1997) [38] showed that the CD 95 ligand/receptor system, a specific mediator of apoptosis is induced by reactive oxygen intermediates. Treatment of Hep G2 cells with bleomycin beside induction of CD 95 induced the production of ROS and in addition GSH depletion. N-acetyl-cysteine, an antioxidant and GSH precursor resulted in partial restoration of intracellular GSH levels and reduced induction of CD 95 ligand mRNA. Induction of CD 95 ligand by bleomycin was further reduced by combined treatment with N-acetyl-cysteine and desferoxamine. The latter leads to complex formation of i.e. iron which might block the formation of ferric peroxide and hydroxyl radicals which are formed via Fe2+-catalyzed Fenton reaction. Cisplatin (cis-Pt) acts via intra- and interstrand cross-links resulting in interference with excission repair and other vital processes and via ROS. During treatment lipidperoxidation occurs and superoxides arise from interaction with DNA [39;40]. Increase of glutathione levels decreases effectiveness of cis-platin [41] and glutathione depletion enhances cytotoxicity of cisplatin [42;43]. Yokomizo and co workers [44] described cisplatin resistent human bladder- and prostatic cancer cells had much higher levels of thioredoxin, a cellular thiol, that functions as a self-defense mechanism in response to oxidative stress. Thioredoxin antisense transfected bladder cancer cells with decreased levels of thioredoxin showed increased sensitivity to cis-Pt and also to other ROS generating agents i.e.doxorubicin, mitomycin C, etoposide, and hydrogen peroxide, as well as UV irradiation. Cis-Pt induced neuropathy is a major limiting factor for treatment. Park and co workers [45] showed that apoptosis of a neuron-neuroblastoma mouse cell line following cis-platin treatment was blocked completely following pretreatment with Nacetylcysteine (precursor of glutathione) and partially after pretreatment with Trolox (vitamin E analogue). Cis-platin-induced p53 accumulation was also blocked (NAC) or retarded (Trolox). The authors suggest that this explains the "neuroprotective" effect of antioxidants during cis-platin treatment. Treatment of rats with cis-platin and GSH prevented neuropathy [46]. The preventive effect of GSH on cis-Pt side effects were confirmed in a human study with 151 patients with ovarian cancer treated with cis-platin

H.K. Biesalski and J. Frank /Antioxidants in Cancer Therapy

257

+/- glutathione [47]. Nephrotoxicity, neurotoxicity and hair loss were significantly reduced in patients receiving six courses of cis-platin. Clinically assessed response to treatment demonstrated a trend towards a better outcome in the GSH+ group, but was not statistically significant. Whether these findings justify a general recommendation to add antioxidants to cis-platin treatment remains questionable.

6. Other Agents Anthracyclines (doxorubicin, daunorubicin) possess the capacity to undergo redox cycling with subsequent ROS generation. There is little doubt that the oxy radical produced by the redox cycling of the quinone/semiquinone metabolite has toxic consequences. This radical is known to cause macro-molecular oxidation. There is, however, an ongoing discussion whether the ROS formed such as superoxides, peroxides and hydroxyl radicals are involved in the cytotoxicity of anthracyclines. The lipid peroxidation observed after therapy with anthracyclines is controversially discussed between a "non major role for free radicals" [48] and the statement from Taatjes and co-workers [49] "induction of oxidative stress is responsible for most if not all biological activity". There is, indeed, evidence from in vitro studies that ROS play an important role. Adriamycin resistant MCF7 human breast cancer cell lines and erythroleukemia cells over express GSTjt, selenium independent GSH and SOD [50]. Doxorubicin (DOX) is widely used in the treatment of cancer, but limited due to the development of a severe form of cardiotoxicity. DOX induced generation of ROS was proposed to be a major mechanism of this side effect. Recently DOX was shown to undergo reductive activation at the reductase domain of the enzyme, endothelial nitric-oxide synthase (eNOS) [51]. The cardivascular toxicity of DOX was attributed to a switch in the encymatic activity of eNOS from a nitric oxide generation enzyme (i.e. eNOS activity) to a superoxide generating enzyme (i.e. NADPH oxidase activity) [51]. In a recent study Kalivendi and co-workers (2001) [52] documented that DOX induces i.e. oxidative stress in endothelial cells. The oxidative stress was inhibited by antisense eNOS oligonucleotide and antioxidant treatment. Antiapoptotic antioxidants (e.g. FeTBAP, ebselen, and a-phenylterf-butyl-nitrone) inhibited DOX induced eNOS transcription. The authors conclude that DOX induced apoptosis of endothelial cells is linked to the redox activation of DOX by eNOS. With respect to the cardiotoxicity antioxidants might be beneficial under in vitro conditions and animal experiments. Clinical trials however failed to demonstrate a beneficial effect of vitamin E during treatment with DOX on side effects including cardiotoxicity [53;54]. Whether antioxidants interfere with the efficacy of tumor treatment has not been elucidated up to now. Involvement of ROS in the cytotoxicity of DOX against cancer cells is probably of importance only in higher doses [55]. Mitomycin C (Naphtoquinon) acts via ROS formation and DNA-adduct formation resulting in intra- and inter-strand cross links and mono- and di-alkylation. There is good evidence that mitomycin C acts via oxidative stress which is strengthened by the observation between cancer cell cytotoxicity and ROS-formation [56] and the relationship between reduction potential and cytotoxicity for different mitomycins. Paclitaxel (Taxol) has multiple effects on tumor cells including DNA fragmentation, inhibition and NO production. In addition, oxidative stress may play an important role since nonataxel enhances the production of phorbol ester induced superoxides by macrophages, an effect which was abolished by prior treatment with SOD and N-acetyl-cysteine, a free radical scavenger [57]. The isoflavonoid Quercetin, antioxidant and general kinase inhibitor, was able to prevent the onset of Taxol-induced cellular detachment in HeLa cells and to protect the cells from cell death. In addition Taxol-induced phosphorylation of p38

258

U.K. Biesalski and J. Frank / Antioxidants in Cancer Therapy

and bc!2 were blocked [58]. Similarity another antioxidant, Resveratrol, present in grapes and red wine, significantly reduces apoptosis in Taxol treated neuroblastoma cells [59]. The anti-estrogens, tamoxifen and diethylstilbesterol exert their genotoxicity either via binding covalently with DNA or generating ROS. Tamoxifen is thought to act primarily by competitive blockade of the estrogen receptor. However, with the emergence of tamoxifen resistant tumor cells these cells are stimulated rather than inhibited resulting in a progression of the disease. As one reason an altered cellular redox status leading to activation of downstream signalling pathways is discussed. Indeed, tamoxifen can induce i.e. dysbalance of the redox status in both directions [60;61]. Changes in the i.e. redox status can lead to the activation of redox-sensitive transcription factors e.g. AP-1. An increase of AP-1 as a consequence of tamoxifen induced oxidative stress is hypothesized to contribute to tamoxifen resistant tumor growth [62]. Schiff and co-workers demonstrated that the development of acquired tamoxifen resistance of xenograft MCF-7 tumors in vivo is associated with both increased susceptibility to oxidative stress and increased AP-1activity. Tamoxifen increases lipid peroxidation and induces the activity of a number of antioxidant enzymes. Prolonged tamoxifen treatment resulted in tamoxifen resistance with increased growth and reduced antioxidant cellular capacity including glutathione depletion. As a consequence of these findings, the question arises at which time point treatment with antioxidants might be beneficial or harmful. If chronic oxidative stress, appearing as a consequence of prolonged tamoxifen treatment, results in activation of JNK. with subsequent increased AP 1 activity, repletion of the i.e. antioxidant capacity via e.g. Nacetyl-cysteine, a GSH precursor, might be an approach. If this will lead to a down regulation of AP-1 and its growth stimulating effect, tamoxifen resistance should be reduced. This, however, has yet not been proven. 7. Alkylating Agents Alkylating agents act by substituting alkyl groups with biological active molecules in particular nucleic acids and cellular enzymes which, as a consequence, are inactivated. Irreversible binding to DNA and alkylating of N7 of guanine results in disruption of DNA function. Beside this effect alkylating agents have also been proposed to form ROS via reduction of an iminium ion to an iminium radical and transfer of the electron to molecular oxygen with subsequent superoxide formation [1;63]. Further reactions leading to the formation of ROS include either the formation of SoxoG via a guanine radical cation or generation of alkyl peroxyl radicals [1;63;64]. Alkylating agents trigger apoptosis through different pathways including ROS formation and ROS defence (Figure 3). Indeed, the antioxidant N-acetyl-cysteine markeldy reduces nitrogen mustard (HN2) induced apoptosis of lymphocytes, even when NAC was added 30 minutes after HN2 administration. The major reason for these effect is the delivery of the substrate for glutathione repletion because alkylating agents deplete glutathione and induce oxidative stress [65]. In addition ROS formation induces heat shock reponse [66]. Consequently the heat shock response due to treatment of cells with alkylating agents is not only a result of protein damage, as proposed [67]. Activation of HSF DNA binding may also occur as a result of depletion of glutathione, oxidative stress and lipid peroxidation [65;66;68]. Liu and coworker (1996) [69] showed an activation of HSF by alkylating agents, triggered by glutathione depletion and oxidation of protein thiols in LLC PK1 cells. Consistent with this suggestion is the fact that DTT (dithiothreitol) prevented the loss of cellular protein thiols and blocked the formation of high molecular weight protein aggregates. Based on the above mentioned experiments it can be argued that the cellular thiol disulfide redox status and formation of disulfide linked aggregates of cellular proteins are linked to HSF1 activation and hsp70

H.K. Biesalski and J. Frank /Antioxidantu in Cancer Therapy

259

transcritpional activation. Restoration or individual differences of i.c glutathione or thiol redox status might reduce the efficacy of the treatment with alkylating agents.

8. Hyperthermia It is well established that the exposure of tumor cells to hyperthermia (HT) above 42°C results in increased cell death [70;71], and it has been hypothesised that hyperthermic injury to cancer cells might be (at least partially) caused by the activation of oxidative stress processes, as documented in human liver cancer [72]and in non-neoplastic cells [73-76]. In liver tissue exposed to hyperthermic perfusion, several markers of oxidative injury have in fact been determined, including the oxidation and depletion of reduced glutathione [74] and the appearance of aldehydic products of lipid peroxidation [72;75;76]. At the cellular level, damage to nucleic acids, proteins, lipids and breakdown of the cellular membrane or lysosomes may be responsible for cell death following HT. At the same time, hyperthermia has a profound effect on the functional and structural integrity of tumor microcirculation [77;78]. Both effects are believed to play a role in the effectiveness of HT in the treatment of tumors. In addition, it has been repeatedly shown that hyperthermia can induce oxidative stress in tissues, due to increased production of reactive oxygen species (ROS) and/or promotion of cellular oxidation events [72-76]. Such an increased formation of ROS during hyperthermia can represent a further major mechanism in cell injury, and could in principle be exploited in hyperthermic treatments of cancer. ROS are in fact cytotoxic, and have been shown to kill cells via apoptosis [19]. Very recently we carried out a couple of experiments to evaluate the role of ROS during HT. The effects of respiratory hyperoxia (RH) and xanthine oxidase (XO) during localized hyperthermia (HT) were investigated by determining markers of oxidative stress. Upon treatment, increases in thiobarbituric acid-reactive substances (marker for lipid peroxidation), protein bound 4-hydroxynonenal (marker for phospholipid peroxidation), protein-associated carbonyl functions (marker for protein oxidation), and number of cells undergoing apoptosis were found in tumor tissue, together with an inhibition of tumor growth [79]. In some studies, the activation of tissue xanthine dehydrogenase to its oxidase form has been implicated as a major mechanism in the oxidative stress caused by hyperthermia [72-74]. Indeed, HT has been shown to induce XO activation [72-74], and it is known that the interaction of XO with its substrate (hypoxanthine) is associated with the generation of superoxide anion as a by-product. One reason for induction of XO is the presence of a so called ischemia/reperfusion in tumor tissues. In an experimental rabbit model [76], hyperthermia substantielly reduced atocopherol content in the tumor. The antitumor effect of hyperthermia was significantly inhibited by the administration of superoxide dismutase and catalase or dimethyl sulfoxide. As long as we do not know, whether antioxidants might weaken this ROS generating effect it seems advisable not to recommend antioxidants during hyperthermia. A further argument against the use of antioxidants during tumor treatment comes from recent studies from Huang and co workers (2001) [80] showing that the treatment of leucemia cells with 2-methoxyestradiol (2MB) resulted in an increased apoptosis. The authors speculated that this was a result of SOD inhibition. Even it is not generally accepted that 2MB is indeed an inhibitor of SOD, 2MB treatment is accompanied by an increased formation of superoxide anion radicals. We showed in vitro that the appropriate administration of the nontoxic "prooxidant" agent 2-ME causes accumulation of cellular ROS which leads to free radical mediated damage to mitochondrial membranes, enhanced lipid peroxidation, elevation in caspase-3 activity, and apoptosis of cancer cell. 2-ME was able to increase intracellular superoxide

260

U.K. Biesalski and J. Frank/Antioxidants in Cancer Therapy

production in our experimental tumor cell line resulting in enhanced cell death. The combination of 2-ME with a further ROS generating treatment (hypoxanthine and xanthine oxidase) enhanced in vitro cytotoxicity. In vivo, 2-ME caused a slight inhibition of tumor growth, with no tumors cured. The combination of 2-ME treatment with localized 44°C hypothermia, respiratory hyperoxia and xanthine oxidase caused a tumor growth delay with 51% of tumors cured. The in vivo results clearly show that a low formation of ROS might be compensated by defence mechanisms of the tumor cells whereas high concentrations of ROS, generated via additional HT, resulted in ROS induced tumor growth delay. As a consequence the maintenance of a dysbalance of low antioxidant defence and high ROS generation (oxidative stress) is proposed to be important to ensure ROS induced cell death of tumor cells.

9. Gene Therapy using Oxidative Stress A novel approach based on an old idea was described by Stegman et al, transferring the gene for the D-amino acid oxidase (DAAO) into glioma cells. DAAO ectopically transfected into the cytoplasme generates HjOi which attacks intracellular organelles and as a hypothesis is not detoxified via catalase present in the peroxisamal matrix. Indeed, exposure to transfected glioma cells to D-alanine resulted in cytotoxicity at concentrations that were non-toxic to parenteral 9L glioma cells. Depletion of cellular glutathione further sensitized 9Ldaaol7 cells to D-alanine (D-Ala). This result, combined with stimulation of pentose phosphate pathway activity and the production of extracellular H2O2 by 9Ldaaol7 cells incubated with D-alanine implicates oxidative stress as the mediator of cytotoxicity [81]. These experiments demonstrate that an increase of i.e. oxidative stress similarly to PDT and HT as described via exogenous administration of a ROS generating substrate might be a novel approach.

10. Conclusion A couple of data exist showing beneficial effects of antioxidants during treatment of cancer cells. However, most of the data are from in vitro studies and their relevance has not been proven yet in clinical trials. In addition, in vitro studies deal with selected antioxidants and selected chemotherapeutica. Despite the high variety chemopreventive drugs which are effective via ROS formation the in vitro results with one compound and one tumor cell line is often taken as a basis to argument that antioxidants might be beneficial during chemotherapy. Most of the treatment strategies have not been evaluated together with antioxidants, even the majority of these strategies is based on ROS formation. If we have evidence that antioxidants either isolated or in combination have a beneficial effect on a specific cancer cell, this should be clarified under in vivo conditions before recommendations are given on theoretical basis. Whether antioxidants may be beneficial or not must be decided case by case. Kind of cancer, art of treatment, individual aspects such as nutritional status, age and gender may influence this approach. Even antioxidants may reduce side effects and improve quality of life, the uncertainty whether antioxidants might be also beneficial for the tumor cells, dealing with antioxidants recommendations should be more critical. A general recommendation of antioxidants as adjuvans during tumor therapy as given by some pharmaceutical companies seems more harmful than a more general recommendation to deal cautious with antioxidants during tumor therapy.

H.K. Biesalski and J. Frank/ Antioxidants in Cancer Therapy

261

References [I] P.Kovacic, J.A.Osuna, Jr. Mechanisms of anti-cancer agents: emphasis on oxidative stress and electron transfer, Curr.Pharm.Des, 6, (2000) 277-309. [2] S.Nair, E.P.Norkus, H.Hertan, C.S.Pitchumoni. Serum and colon mucosa micronutrient antioxidants: differences between adenomatous polyp patients and controls, Am.J.Gastroenterol., 96, (2001) 3400-3405. [3] G.Maiani, G.Pappalardo, A.Ferro-Luzzi, A.Raguzzini, E.Azzini, A.Guadalaxara, M.Trifero, T.Frommel, S.Mobarhan. Accumulation of beta-carotene in normal colorectal mucosa and colonic neoplastic lesions in humans, Nutr.Cancer, 24, (1995) 23-31. [4] M.E.Begin. Effects of polyunsaturated fatty acids and of their oxidation products on cell survival, Chem.Phys.Lipids, 45, (1987) 269-313. [5] M.E.Begin, G.EIIs, D.F.Horrobin. Polyunsaturated fatty acid-induced cytotoxicity against tumor cells and its relationship to lipid peroxidation, J.Natl.Cancer Inst., 80, (1988) 188-194. [6] U.N.Das, M.E.Begin, G.EIIs, Y.S.Huang, D.F.Horrobin. Polyunsaturated fatty acids augment free radical generation in tumor cells in vitro, Biochem.Biophys.Res.Commun., 145, (1987) 15-24. [7] C.Lhuillery, S.Cognault, E.Germain, M.LJourdan, P.Bougnoux. Suppression of the promoter effect of polyunsaturated fatty acids by the absence of dietary vitamin E in experimental mammary carcinoma, Cancer Lett., 114, (1997) 233-234. [8] LW.Oberley, G.R.Buettner. Role of superoxide dismutase in cancer: a review, Cancer Res., 39, (1979) 1141-1149. [9] LW.Oberley, T.D.Obertey. Role of antioxidant enzymes in cell immortalization and transformation, Mol.Cell Biochem., 84, (1988) 147-153. [10] L.W.Oberley. Anticancer therapy by overexpression of superoxide dismutase, Antioxid.Redox.Signal., 3, (2001)461-472. [II] S.X.Skapek, O.M.Colvin, O.W.Griffith, D.R.Groothuis, E.V.Colapinto, Y.Lee, J.Hilton, G.B.EIion, D.D.Signer, H.S.Friedman. Buthionine sulfoximine-mediated depletion of glutathione in intracranial human glioma-derived xenografls, Biochem. Pharmacol., 37, (1988) 4313-4317. [12] D.Labriota, R.Livingston. Possible interactions between dietary antioxidants and chemotherapy, Oncology (Huntingt), 13, (1999) 1003-1008. [13] R.A.Gatenby, H.B.Kessler, J.S.Rosenblum, L.R.Coia, P.J.Moldofsky, W.H.Hartz, G.J.Broder. Oxygen distribution in squamous cell carcinoma metastases and its relationship to outcome of radiation therapy, Int.J.Radiat.Oncol.Biol.Phys., 14, (1988)831-838. [14] K.A.Kennedy. Hypoxic cells as specific drug targets for chemotherapy, Anticancer Drug Des, 2, (1987) 181-194. [15] M.Hockel, C.Knoop, K.Schlenger, B.Vorndran, E.Baussmann, M.Mrtze, P.G.Knapstein, P.Vaupel. Intratumoral pO2 predicts survival in advanced cancer of the uterine cervix, Radiother.Oncol., 26, (1993) 45-50. [16] M.Nordsmark, M.Overgaard, J.Overgaard. Pretreatment oxygenation predicts radiation response in advanced squamous cell carcinoma of the head and neck, Radiother.Oncol., 41, (1996) 31-39. [17] J.S.Daniels, K.S.Gates, C.Tranche, M.M.Greenberg. Direct evidence for bimodal DNA damage induced by tirapazamine, Chem.Res.Toxicol., 11, (1998) 1254-1257. [18] W.E.Wung, S.B.Howell. Hypoxanthine concentrations in normal subjects and patients with solid tumors and leukemia, Cancer Res., 44, (1984) 3144-3148. [19] T.M.Buttke, P.A.Sandstrom. Oxidative stress as a mediator of apoptosis, Immunol.Today, 15, (1994) 7-10. [20] G.Minotti. Sources and role of iron in lipid peroxidation, Chem.Res.Toxicol., 6, (1993) 134-146. [21] R.Roots, S.Okada. Estimation of life times and diffusion distances of radicals involved in x-ray-induced DNA strand breaks of killing of mammalian cells, Radiat.Res., 64, (1975) 306-320. [22] A.T.al Kazwini, P.O'Neill, G.E.Adams, E.M.Fielden. Radiation-induced energy migration within solid DNA: the role of misonidazole as an electron trap, Radiat.Res., 121, (1990) 149-153. [23] V.Darley-Usmar, H.Wiseman, B.Halliwell. Nitric oxide and oxygen radicals: a question of balance, FEBS Lett., 369, (1995)131-135. [24] C.B.Brezden, R.A.McClelland, A.M.Rauth. Mechanism of the selective hypoxic cytotoxicity of 1-methyl-2nitroimidazole, Biochem.Pharmacol., 48, (1994)361-370.

262

H.K. Biesalski and J. Frank / Antioxidants in Cancer Therapy

[25] H.Sies. Biochemistry of oxidative stress. Angew.Chemie IntEdition 25,1058-1071.1986. [26] A.H.Nias.

Radiation and platinum drug interaction, IntJ.Radiat.Bfol.Relat Stud.Phys.Chem.Med., 48,

(1985)297-314. [27] S.R.Chatterjee, H.Possel, T.S.Srivastava, J.P.Kamat, G.Wolf, T.P.Devasagayam. Photodynamic effects induced by meso-tetrakis[4-(carboxymethyleneoxy)phenyQ porphyrin on isolated Sarcoma 180 astites mitochondria, J.Photochem.Photobiol.B, 50, (1999) 79-87. [28]

C.Salet,

G.Moreno.

Photosensitization

of

mitochondria.

Molecular

and

cellular

aspects,

J.Photochem.Photobiol.B, 5, (1990) 133-150. [29] J.R.Shutok, M.H.Wade, C.W.Lin. Subcellular localization of hematoporphyrin derivative in bladder tumor cells in culture, Photochem.Photobiol., 51, (1990) 451-457. [30] M.C.Luna, A.Ferrario, S.Wong, A.M.Fisher, C.J.Gomer. Photodynamic therapy-mediated oxidative stress as a molecular switch for the temporal expression of genes Ikjated to the human heat shock promoter, Cancer Res., 60, (2000) 1637-1644. [31] L.Ftohe, R.Brigelius-Ftohe, C.Saliou, M.G.Traber, L.Packer

Redox regulation of NF-kappa B activation.

Free Radic.Biol.Med., 22, (1997) 1115-1126. [32] C.K.Sen, L.Packer. Antioxidant and redox regulation of gene transcription, FASEB J., 10, (1996) 709-720. [33] S.Grether-Beck, S.OIaizola-Hom, H.Schmitt, M.Grewe, A.Jahnke, J.P.Johnson,

K.Briviba, H.Sies,

J.Krutmann. Activation of transcription factor AP-2 mediates UVA radiation- and singlet oxygen-induced expression of the human intercellular adhesion molecule 1 gene, Proc.Nati.Acad.Sci.U.S.A. 93, (1996) 14586-14591. [34] LO.Ktotz, K-Briviba, H.Sies. Singlet oxygen mediates the activation of JNK by UVA radiation in human skin fibroblasts, FEBS Lett.. 408, (1997) 289-291. [35] J.Tao, J.S.Sanghera, S.LPelech, G.Wong, J.G.Levy. Stimulation of stress-activated protein kinase and p38 HOG1 kinase in murine keratinocytes following photodynamic therapy with benzoporphyrin derivative. J.Btol.Chem., 271, (1996) 27107-27115. [36] S.M.Keyse.

An emerging family of dual specificity MAP kinase phosphatases, Bkxhim.Biopnys Acta,

1265, (1995)152-160. [37] K.Scharffetter-Kochanek, M.WIaschek, P.Brenneisen, M.Schauen, R.BIaudschun, J.Wenk.

UV-induced

reactive oxygen species in photocarcinogenesis and photoaging, Biol.Chem., 378, (1997) 1247-1257. [38] H.Hug, S.Strand, A.Grambihler, J.Galte, V.Hack, W.Stremmel, P.H.Krammer, P.R.Galle. Reactive oxygen intermediates are involved in the induction of CD95 ligand mRNA expression by cytostatic drugs in hepatoma cells, J.Biol.Chem., 272, (1997) 28191-28193. [39] H.Masuda, T.Tanaka, U.Takahama. Cisplatin generates superoxide anion by interaction with DMA in a cell-free system, Biochem.Biophys.Res.Commun., 203, (1994) 1175-1180. [40] R.Uslu, B.BonavKJa. Involvement of the mitochondrion respiratory chain in the synergy achieved by treatment of human ovarian carcinoma cell lines with both tumor necrosis factor-alpha and cisdiamminedichtoroplatinum, Cancer, 77, (1996) 725-732. [41] S.A.Shellard, LK.Hosking, B.T.Hill. Anomalous relationship between cisplatin sensitivity and the formation and removal of platinum-DNA adducts in two human ovarian carcinoma cell lines in vitro, Cancer Res., 51, (1991)4557-4564. [42] A.Miyajima, J.Nakashima, K.Yoshioka, M.Tachibana, H.Tazaki, M.Murai. Role of reactive oxygen species in cis-dichlorodiammineplatinum-induced cytotoxicity on bladder cancer cells, Br.J.Cancer, 76, (1997) 206210.

[43] A.Miyajima, J.Nakashima, M.Tachibana, K.Nakamura, M.Hayakawa, M.Murai. N-acetyteysteine modifies cis-dichtorodiammineplatinum-induced effects in bladder cancer cells, Jpn.J.Cancer Res., 90. (1999) 565570.

[44] A.Yokomizo, M.Ono, H.Nanri, Y.Makino, T.Ohga, M.Wada, T.Okamoto, J.Yodoi, M.Kuwano, K.Kohno. Cellular levels of thioredoxin associated with drug sensitivity to cisplatin, mitomvcin C, doxorubicin, and etoposide, Cancer Res., 55, (1995) 4293-4296. [45] S.A.Park, K.S.Choi, J.H.Bang, K.Huh, S.U.Kim.

Cispiatin-induced apoptotic cell death in mouse hybrid

neurons is blocked by antioxidants through suppression of cisplatin-mediated accumulation of p53 but not of Fas/Fas ligand, J.Neurochem., 75, (2000) 946-953.

H.K. Biesalski and J. Frank / Antioxidants in Cancer Therapy

263

[46] F.P.Hamers, J.H.Brakkee, E.Cavalletti, M.Tedeschi, LMarmontt, G.Pezzoni, J.P.Neijt, W.H.Gispen. Reduced glutathione protects against cisplatin-induced neurotoxicity in rats, Cancer Res., 53, (1993) 544549. [47] J.F.Smyth, A.Bowman, T.Perren, P.Wilkinson, R.J.Prescott, K.J.Quinn, M.Tedeschi. Glutathione reduces the toxicity and improves quality of life of women diagnosed with ovarian cancer treated with cisplatin: results of a double-blind, randomised trial, Ann.Oncol., 8, (1997) 569-573. [48] J.Tarasiuk, K.Tkaczyk-Gobis, B.Stefanska, M.Dzieduszycka, W.Priebe, S.Martelli, E.Borowski. The role of structural factors of anthraquinone compounds and their quinone-modified analogues in NADH dehydrogenase-catalysed oxygen radical formation, Anticancer Drug Des, 13, (1998) 923-939. [49] D.J.Taatjes, D.J.Fenick, G.Gaudiano, T.H.Koch. A redox pathway leading to the alkylation of nucleic acids by doxorubicin and related anthracyciines: application to the design of antitumor drugs for resistant cancer, Curr.Pharm.Des, 4, (1998) 203-218. [50] J.C.Schisselbauer, M.Crescimanno, N.D'Alessandro, M.CIapper, S.Toulmond, H.Tapiero, K.D.Tew. Glutathione, glutathione S-transferases, and related redox enzymes in Adriamycin-resistant cell lines with a multidrug resistant phenotype, Cancer Commun., 1, (1989) 133-139. [51] J.Vasquez-Vivar, P.Martasek, N.Hogg, B.S.Masters, K.A.Pritchard, Jr., B.Kalyanaraman. Endothelial nitric oxide synthase-dependent superoxide generation from adriamycin. Biochemistry, 36, (1997) 11293-11297. [52] S.V.Kalivendi, S.Kotamraju, H.Zhao, J.Joseph, B.Kalyanaraman. Doxorubicin-induced apoptosis is associated with increased transcription of endothelial nitric-oxide synthase. Effect of antiapoptotic antioxidants and calcium, J.Biol.Chem., 276, (2001) 47266-47276. [53] S.S.Legha, Y.M.Wang, B.Mackay, M.Ewer, G.N.Hortobagyi, R.S.Benjamin, M.K.AIi. Clinical and pharmacoiogic investigation of the effects of alpha-tocopherol on adriamycin cardiotoxicity, Ann.N.Y.Acad.ScL, 393:411-8., (1982)411-418. [54] S.Weitzman, B.Lorell, R.Carey. Prospective study of tocopherol prophylaxis for anthracycline cardiac toxicity. Curr. Ther. Res. 28, (1980) 682-686. [55] H.G.Keizer, H.M.Pinedo, G.J.Schuurhuis, H.Joenje. Doxorubicin (adriamycin): a critical review of free radical-dependent mechanisms of cytotoxicity, Pharmacol.Ther., 47, (1990) 219-231. [56] B.K.Sinha. Free radicals in anticancer drug pharmacology, Chem.Biol.Interact., 69, (1989) 293-317. [57] H.O.Pae, C.D.Jun, J.C.Yoo, H.J.Kwak, S.J.Lee, Y.A.Kook, R.K.Park, H.T.Chung. Enhancing and priming of macrophages for superoxide anion production by taxol, Immunopharmacol.lmmunotoxicol., 20, (1998) 27-37. [58] M.Marone, G.D'Andrilli, N.Das, C.Ferlini, S.Chatterjee, G.Scambia. Quercetin abrogates taxol-mediated signaling by inhibiting multiple kinases, Exp.Cell Res., 270, (2001) 1-12. [59] G.Nicolini, R.Rigolio, M.Miloso, A.A.Bertelli, G.Tredici. Anti-apoptotic effect of trans-resveratrol on paclitaxel-induced apoptosis in the human neuroblastoma SH-SY5Y cell line, Neurosci.Lett., 302, (2001) 41-44. [60] S.S.Dehal, D.Kupfer. Evidence that the catechol 3,4-Dihydroxytamoxifen is a proximate intermediate to the reactive species binding covalently to proteins, Cancer Res., 56, (1996) 1283-1290. [61] H.Wei, Q.Cai, L.Tian, M.Lebwohl. Tamoxifen reduces endogenous and UV light-induced oxidative damage to DNA, lipid and protein in vitro and in vivo, Carcinogenesis, 19, (1998) 1013-1018. [62] R.Schiff, P.Reddy, M.Ahotupa, E.Coronado-Heinsohn, M.Grim, S.G.Hilsenbeck, R.Lawrence, S.Deneke, R.Herrera, G.C.Chamness, S.A.Fuqua, P.H.Brown, C.K.Osborne. Oxidative stress and AP-1 activity in tamoxtfen-resistant breast tumors in vivo, J.Natl.Cancer Inst., 92, (2000) 1926-1934. [63] K.Mizumoto, P.A.GIascott, Jr., J.L.Farber. Roles for oxidative stress and poly(ADP-ribosyl)ation in the killing of cultured hepatocytes by methyl methanesulfonate, Biochem.Pharmacol., 46, (1993) 1811-1818. [64] P.J.Dandliker, R.E.Holmlin, J.K.Barton. Oxidative thymine dimer repair in the DNA helix, Science, 275, (1997)1465-1468. [65] Q.Chen, J.LStevens. Inhibition of iodoacetamide and t-butylhydroperoxide toxicity in LLC-PK1 cells by antioxidants: a role for lipid peroxidation in alkylation induced cytotoxicity, Arch.Biochem.Biophys., 284, (1991)422-430. [66] M.LFreeman, M.J.Borrelli, K.Syed, G.Senisterra, D.M.Stafford, J.R.Lepock. Characterization of a signal generated by oxidation of protein thiols that activates the heat shock transcription factor, J.Cell Physiol, 164, (1995)356-366.

264

H.K. Biesalski and J. Frank / Antioxidants in Cancer Therapy

[67] R.A.Kroes, K.Abravaya, J.SekJenfekJ, R.I.Morimoto.

Selective activation of human heat shock gene

transcription by nitrosourea antitumor drugs mediated by isocyanate-induced damage and activation of heat shock transcription factor, Proc.Natl.Acad.Sci.U.S.A, 88, (1991) 4825-4829. [68] P.AItevi, M.Anastasia, F.Cajone, P.Ciuffreda, A.M.Sanvito. Structural requirements of aldehydes produced in LPO for the activation of the heat-shock genes in HeLa cells, Free Radic.Biol.Med., 18, (1995) 107-116. [69] H.Liu, R.Lightfoot, J.L.Stevens. Activation of heat shock factor by alkylating agents is triggered by glutathione depletion and oxidation of protein thiols, J.Biol.Cnem., 271, (1996) 4805-4812. [70] W.C.Dewey, LE.Hopwood, S.A.Sapareto, LE.Gerweck.

CeHular responses to combinations of

hyperthermia and radiation, Radiology, 123, (1977) 463-474. [71] C.Streffer. Molecular and cellular mechanisms of hyperthermia, in: M.H.Segenschmiedt, M.Fessenden, C.C.Vemon (Eds.), Thermoradiotherapy and Therrnochemotherapy, Springer-Veriag, Berlin, 1995, pp. 4774. [72] J.LSkibba, E.J.Quebbeman, J.H.Kafofleisch.

Nitrogen metabolism and lipid peroxidation during

hyperthermic perfuston of human livers with cancer, Cancer Res., 46, (1986) 6000-6003. [73] P.S.Lin, S.Quamo, K.C.Ho, J.Gladding. Hyperthermia enhances the cytotoxic effects of reactive oxygen species to Chinese hamster cells and bovine endothelial cells in vitro, Radiat.Res., 126, (1991)43-51. [74] R.H.Powers, A.Stadnicka, J.H.Kalbfleish, J.LSkibba. Involvement of xanthine oxidase in oxidative stress and iron release during hyperthermic rat liver perfusion, Cancer Res., 52, (1992) 1699-1703. [75] J.LSkibba, E.A.Gwartney. Liver hyperthermia and oxidative stress: role of iron and aldehyde production, Int.J.Hyperthermia, 13, (1997) 215-226. [76] T.Yoshikawa, S.Kokura, K.Tainaka, K.ltani, H.Oyamada, T.Kaneko, Y.Naito, M.Kondo. The rote of active oxygen species and lipid peroxidation in the antitumor effect of hyperthermia, Cancer Res., 53, (1993) 2326-2329. [77] D.K.Kelleher, P.Vaupel. Tumor microenvironment: characterization and significance for hyperthermia treatment. Exptl.Oncol. 17, 256-268.1995. [78] P.Vaupel. Pathophysiological mechanisms of hyperthermia in cancer therapy, in: M.Gautherie (Ed.), Biological Basis of Oncologic Thermotherapy, Springer-Veriag, Berlin, 1990, pp. 73-134. [79] J.Frank, D.K.Kelleher, A.Pompella, O.Thews, H.K.Biesalski, P.Vaupel.

Enhancement of oxidative cell

injury and antitumor effects of localized 44 degrees C hyperthermia upon combination with respiratory hyperoxia and xanthine oxidase, Cancer Res., 58, (1998) 2693-2698. [80] P.Huang, LFeng, E.A.OIdham, M.J.Keating, W.PIunkett.

Superoxide dismutase as a target for the

selective killing of cancer cells, Nature, 407, (2000) 390-395. [81] LD.Stegman, H.Zheng, E.R.Neal, O.Ben Yoseph, LPollegioni, M.S.Pilone, B.D.Ross.

Induction of

cytotoxic oxidative stress by D-alanine in brain tumor cells expressing Rhodotorula gracilis D-amino acid oxidase: a cancer gene therapy strategy, Hum.Gene Ther., 9, (1998) 185-193.

Thiol Metabolism and Redox Regulation of Cellular Functions A. Pompella et at. (Eds.) IOS Press, 2002

265

Bisulfide Exchange in CD4 Lisa J. MATTHIAS, Patricia T. W. YAM, Xing-Mai JIANG and Philip J. HOGG* Centre for Thrombosis and Vascular Research, School of Medical Sciences, University of New South Wales, NSW 2052, Australia *Tel: +61-2-9385-1004; Fax: +61-2-9385-1389; E-mail: [email protected]

1. The Bisulfide Bond It is generally considered that disulfide-bonds have been added during evolution to enhance the stability of proteins which must function in a fluctuating cellular environment. Bisulfide bonds have different effects on the unfolded versus the folded state of proteins. Disulfide bonds lower the entropy of the unfolded form, making it less favourable compared to the folded form [1-3]. In other words, disulfide bonds stabilise the native conformation of a protein by destabilising the unfolded form. On the other hand, disulfide bonds can decrease the stability of folded proteins due to bond enthalpy effects and strain. A method for calculating the strain of a disulfide-bond uses dihedral angles. A dihedral angle is the angle of rotation about a certain bond. The dihedral strain energy of a disulfide-bond can be calculated from the five dihedral angles of a disulfide-bond [4, 5]. The calculated strain energies only consider the dihedral angles of the disulfide-bond and do not include other factors such as bond lengths, bond angles, and van der Waals contacts in calculating energy. The findings of Pjura et al. [6], however, indicate that such calculations can give useful semi-quantitative insights into the amount of strain in a disulfide-bond. There is a correlation between the redox potentials of disulfide-bonds and their calculated dihedral energies, which is of relevance to this study. Those bonds with higher dihedral energies are more easily reduced [3, 5, 7]. Cytosolic proteins rarely contain disulfide-bonds, presumably because the intracellular environment is too reducing to allow cysteine oxidation to the disulfide [8]. In contrast, secreted proteins usually contain disulfide-bonds. The prevailing view is that disulfide-bonds in secreted proteins are inert because of the oxidizing nature of the extracellular milieu. We have suggested that this is not necessarily the case and that certain secreted proteins contain one or more disulfide-bonds that can exchange and that this exchange is central for the function of the protein [9, 10]. This article discusses disulfidebond exchange in the T cell receptor, CD4.

2.CD4 CD4 is expressed on most thymocytes and on the subset of peripheral T lymphocytes that includes helper T cells and binds to class II MHC to enhance the T cell response. Another ligand for CD4 is the human immunodeficiency virus (HIV-1). HIV-1 binds to CD4 and one of several chemokine receptors which triggers fusion of the viral and cell membranes. CD4 has a molecular mass of 55 kDa and consists of an extracellular portion (residues 1371), a transmembrane segment (372-395), and a cytoplasmic tail (396-433). The extracellular portion consists of four immunoglobulin (Ig)-like domains, Dl to D4 [11-14]. Class II MHC binding extends over Dl and D2, while HIV-1 gp!20 binds to Dl.

266

L.J. Matthias et al. / Disulfide Exchange in CD4

3. CD4 Disulfide Bonds The backbone of Ig domains are defined by (5 strands indicated by letters A through G. Following standard Ig nomenclature, one (5 sheet of the sandwich-like structure contains strands A, B and E, and the other sheet has strands C, C, F and G. The disulfide-bond in Dl of CD4 is between Cysl6 in strand B and Cys84 in strand F [11-13]. Similarly, the disulfide-bond in D4 is between Cys303 in strand B and Cys345 in strand F [11, 14]. In contrast, D2 has a truncated (5 barrel (75 residues compared to -100 residues) and a nonstandard disulfide-bond between Cysl30 in strand C and Cysl59 in strand F [11-13]. In effect, D2 has lost the Cys in strand B and acquired one in strand C. This means that the disulfide-bond in D2 is between strands in the same sheet rather than between sheets as is usual. The Cys in strand C eliminates the normally conserved Trp in that position [11-13]. The geometry and strain of the D2 disulfide-bond is also unusual. The disulfide is right- rather than left-handed and has a short (3.92 A) Ca-Ca distance compared to standard Ig disulfides [5] (6.6 A) and right-handed disulfides [5] (5.07 A). The D2 disulfide-bond has a high strain energy (4.74 kcal-mol"1) compared to the Dl (2.28 kcal-mol"1) and D4 (1.71 kcal.mol"1) disulfide-bonds. Notably, the arrangement of the dihedral angles of the D2 disulfide (data not shown) place it in the family of disulfide-bonds (15) that bridge between antiparallel ^-strands rather in the family of Ig disulfides. In addition, the D2 disulfide makes the least contribution to overall stability from enthalpy calculations [2] (D2 disulfide is 3.65 kcal.mol"1 compared to the Dl, 4.36 kcal.mol"1, and D4, 3.97 kcal.mol"1, disulfides). The characteristics of the CD4 disulfides are summarized in Table 1.

Table 1. Characteristics of the CD4 dlsulflde-bonds Domain disulfide-bond

C -C bond length, A

Dihedral Strain Energy, kcal.mor1

Number of residues in disulfide crosslink

Change in enthalpy, kcal.mol1

D1 D2 D4

6.58 3.92 6.56

2.28 4.74 1.71

67 30 43

4.36 3.65 3.97

These features led us to hypothesize that the D2 disulfide-bond was redox active on the cell-surface, that is it could exist in either the oxidised or reduced dithiol forms. Indirect support for this theory was provided by the finding that thiol oxidizing reagents such as HgCb [16] and 5,5'-dithiobis(2-nitrobenzoic acid) [17] facilitate dimerization and oligomerization of CD4 on the cell-surface and that uptake of HIV-1 by CD4+ T cells is inhibited by membrane-impermeable thiol-reactive reagents [18]. We have found that the D2 disulfide-bond can exist in the reduced dithiol form on the cell-surface.

4. Cell-Surface CD4 contains Free Thiol(s) We labeled T cells with either sulfosuccinimidobiotin (SSB) or 3-(N-maleimidylpropionyl)biocytin (MPB). Both reagents are substantially membrane-impermeable [19, 20]. SSB labels primary amines in CD4 while MPB will only label CD4 containing free

L.J. Matthias el al. / Disulfide Exchange in CD4

267

thiols. There are no unpaired cysteines in the extracellular part of CD4 so incorporation of MPB would indicate reduction of one or more of the three disulfide-bonds. The human T cell line, CEM-T4, was incubated with either SSB or MPB and the biotin-labeled proteins were collected on streptavidin agarose, resolved on SDS-PAGE and blotted with the CD4 monoclonal antibody, Leu3a. Cell-surface CD4 incorporated both SSB and MPB (Fig. 1 A). Labeling with MPB was thiol specific as pre-blocking of the MPB with GSH ablated labeling. Comparison of the amount of SSB- versus MPB-labeled CD4 indicated that ~40% of the CD4 on the surface of CEM-T4 cells contained one or more free thiols (Fig. 1A). To confirm the incorporation of MPB into CD4 the receptor was immunoprecipitated from MPB-labeled CEM-T4 cells and the biotin label was detected by blotting with streptavidin-peroxidase. MPB-labeled CD4 is shown in Fig. IB. Labeling was again thiol specific as pre-blocking of the MPB negated incorporation.

^./

Fig. 1. Cell-surface CD4 contains free thlol(s). A, CEM-T4 cells were labeled with either SSB or MPB. The biotin-labeled proteins were collected on streptavidin-agarose beads, resolved on SDS-polyacrylamide gel electrophoresis (SOS-PAGE) and Western blotted using the CD4 monoclonal antibody, LeuSa. Lane 1 is CEM-T4 lysate (from 106 cells), lane 2 is SSB-labeled CEM-T4 CD4, while lane 3 is MPB-labeled CEM-T4 CD4. Biotin-labeled proteins were from 5 x 106 cells. Lane 4 is a control experiment where MPB was pre-blocked with reduced glutathione (GSH) prior to incubation with CEM-T4 cells. The positions of Mr markers in kDa are shown at left.

B, CEM-T4 cells were labeled with MPB and the CD4 immunoprecipitated with LeuSa

monoclonal antibody and goat anti-mouse IgG-coated magnetic microbeads.

The CD4 was

resolved on SDS-PAGE and blotted with streptavidin peroxidase to detect the biotin label. Lane 1 is MPB-labeled CD4 (from 7 x 106 cells). Lane 2 is a control experiment where MPB was preblocked with GSH prior to incubation with CEM-T4 cells.

CD4 on the surface of human blood T cells and monocyte/macrophages and the human monocyte/macrophage line, THP-1, also incorporated MPB (not shown). Activation of blood T cells with phytohemagglutinin for 3 days caused -45% increase in the fraction of total cell-surface CD4 that was labeled with MPB (not shown). Another Ig fold receptor, Thy-1, was not labeled with MPB on the TIB-47 cell line (not shown).

268

L.J. Matthias et al. / Disulfide Exchange in CD4

The thiol specificity of MPB was confirmed by labeling purified plasma proteins that lack or contain a free thiol. MPB labeled serum albumin, which contains a free thiol, but not prothrombin, which does not contain free thiols (not shown).

5. The Domain 2 Disulfide of CD4 is Redox Active These labeling experiments demonstrated that one or more of the 3 disulfide-bonds in cellsurface CD4 could exist in the reduced dithiol form. To determine which of the disulfidebonds was redox active, we stably transfected disulfide-bond mutants of CD4 into cultured cells and tested for the presence of free thiols by labeling with MPB. We prepared Cys to Ala mutants of the three pairs of Cys residues in domains 1, 2 or 4 of CD4. Several attempts at stably transfecting the human CD4- T cell line, A2.01, with the CD4 disulfidebond mutants in the eukaryotic expression vector SRawere unsuccessful (not shown). Stable transfection of the equivalent Cys to Ser mutants into A2.01 cells was also unsuccessful (not shown). In contrast, wild type CD4 was expressed on the surface of A2.01 cells (not shown). These findings suggested that the disulfide-bond mutants were being recognized by the T cells as mis-folded and targeted for degradation. Transfection and surface expression of the CD4 disulfide-bond mutants was achieved, however, in human fibrosarcoma cells. Surface expression of wild-type CD4 and the D2 disulflde mutant in HT1080 cells was confirmed by flow cytometry (not shown). Western blot of the wild-type cell lysate indicated the presence of both monomers and dimers of CD4 (Fig. 2A). The CD4 dimers resolved into monomers upon reduction and alkylation of the lysate (not shown), which indicated that the dimers were disulfide-bonded. The extent of dimer formation varied with experiment but was always apparent. Dimers of the CD4 Dl and D4 disulflde mutants also formed (not shown). In contrast, the CD4 D2 disulflde mutant did not form dimers (Figs. 2A). Labeling of wild-type CD4 and the D2 disulflde mutant on the surface of HT1080 cells with MPB is shown in Fig. 2B. The monomeric forms of wild-type CD4 (Fig. 2B) and the Dl and D4 disulfide-bond mutants (not shown) incorporated MPB whereas the D2 disulflde mutant was not labeled (Fig. 2B), despite approximately equivalent cell-surface expression of the monomeric forms of the wild-type and mutant CD4's. This result indicated that the D2 disulflde of CD4 could exist in the reduced dithiol form on the cellsurface and that it was involved in thiol-dependent dimerization of CD4.

6. Reduction of Cell-Surface and Soluble CD4 by Thioredoxin The finding that T cell activation resulted in reduction of cell-surface CD4 suggested that the oxidation state of the D2 disulfide-bond was controlled by the cell. This could have been accomplished by a secreted disulfide-bond reductase. Incubation of CEM-T4 cells with increasing concentrations of the protein reductant, thioredoxin, resulted in increasing reduction of cell-surface CD4 (Fig. 3 A). The redox properties of thioredoxin were required for reduction of cell-surface CD4 as a redox inactive thioredoxin mutant [21] did not reduce CD4 (Fig. 3B). Incubation of CEM-T4 cells with the same concentrations of another disulfide-bond reductase, protein disulflde isomerase, did not result in increased labeling of CD4 with MPB (not shown). To determine if CD4 was a substrate for thioredoxin independent of the cellsurface we tested whether thioredoxin could reduce a recombinant soluble fragment of CD4 containing the 4 Ig-like domains (sCD4) [22]. Reduction of sCD4 by different

L.J. Matthias et al. /Bisulfide Exchange in CD4

269

concentrations of thioredoxin and incorporation of MPB into reduced sCD4 and thioredoxin is shown in Fig. 3.

<+MPB-CD4

26-

Fig. 2. The domain 2 disulfide of CD4 is redox active. A, The HT1080 CD4 wt (lane 1) and CD4 D2mutant (lane 2) cells were lysed and lysate from 2 x 10s cells was resolved on SDS-PAGE and Western blotted using Leu3a. M represents CD4 monomer while D represents CD4 dimer. B, The HT1080 CD4 wt (lane 1) and CD4 D2mutant (lane 2) cells were labeled with MPB and the MPB-labeled proteins were collected on streptavidin-agarose beads, resolved on SDS-PAGE and Western blotted using Leu3a. MPB-labeled proteins were from 7 x 10s cells. The positions of Mr markers in kDa are shown at left.

•4-CD4 •4-CD4 •«-Trx

Fig. 3. Reduction of cell-surface and soluble CD4 by thioredoxin.

A, CEM-T4 cells were

incubated without (lane 2) or with increasing concentrations of human thioredoxin (lanes 3-5) and then labeled with MPB. The biotin-labeled proteins were resolved on SDS-PAGE and Western blotted using Leu3a monoclonal antibody. Lane 1 is CEM-T4 lysate. The positions of Mr markers in kDa are shown at left. B, CEM-T4 cells were incubated with either redox active thioredoxin mutant (lane 2) or redox inactive thioredoxin mutant (lane 3) and then labeled with MPB. The MPB-labeled proteins were processed as described in part A. Lane 1 is untreated CEM-T4 cells. The positions of Mr markers in kDa are shown at left. C, sCD4 was incubated without (lane 1) or with (lanes 2 and 3) thioredoxin or dithiothreitol (lane 4) and then labeled with MPB. The sCD4 was resolved on SDS-PAGE and blotted with streptavidin peroxidase to detect the biotin label. MPB-labeled sCD4 and thioredoxin are indicated. The positions of Mr markers in kDa are shown at left.

270

LJ. Matthias et al. / Disulfide Exchange in CD4

The reduction was selective for thioredoxin as equivalent concentrations of the strong reductant dithiothreitol (Fig. 3C) or protein disulfide isomerase (not shown) did not reduce sCD4. Soluble gp!20, the HIV-1 envelope glycoprotein that binds CD4, was also tested as a substrate for thioredoxin. Neither thioredoxin nor dithiothreitol reduced gp!20 measured by labeling with MPB. 7. Conclusions We show herein that the CD4 D2 disulfide-bond exists in either the oxidised or reduced dithiol form on the cell-surface. This is the first example of disulfide exchange in an Ig fold. Bisulfides between hydrogen-bonded residues which bridge across the nearest neighbor positions in antiparallel ^-strands are uncommon [23], although crystal structures of model cystine peptides [24] and antiparallel dimers [25] have demonstrated that disulfide bridging between strands in antiparallel (J-sheets can occur. Side-by-side disulfide-bonds have been predicted in other members of the Ig superfamily [26]. Interestingly, the Ig domain of the a chain of CDS, a receptor with a similar function to CD4, also has an unconventional disulfide linkage [27]. There are three Cys in this Ig domain, two in the usual locations and a third two residues before the conserved Tip in strand C. The disulfide in this domain links this extra Cys in strand C with the Cys in strand B, leaving the one in stand F unpaired [27]. The observation that activation of blood T cells resulted in an increase in the dithiol form of cell-surface CD4 suggested that the cells secreted a reductase which reduced the D2 disulfide. Thioredoxin is a 12 kDa dithiol/disulfide redox protein [28] that is secreted by CD4+ T lymphocytes and binds to the cell-surface [29-32]. T cell activation is associated with increased secretion of thioredoxin [29, 31]. Thioredoxin reduced cellsurface and soluble CD4, but not soluble gp!20. Equivalent concentrations of protein disulfide isomerase and dithiothreitol did not reduce CD4.

CD4

Rg. 4. Model for disulfide exchange In domain 2 of CD4. Trx is thioredoxin. The thioredoxin dithiol/disulfide represents the active site cysteines in the sequence CysGlyProCys [28].

L.J. Matthias et al. / Disulfide Exchange in CD4

271

These findings suggested that the oxidised and reduced forms of CD4 exist on the cellsurface in equilibrium which is controlled by thioredoxin secreted by the cell. A model of these events is shown in Fig. 4. The model predicts that reduction of the D2 disulfide by thioredoxin will result in oxidation of the active-site dithiol of thioredoxin. For thioredoxin to catalyze reduction of another CD4 molecule a mechanism is required to reduce the oxidised form of the protein. This may be accomplished by thioredoxin reductase which is secreted by peripheral blood cells and is in plasma at a concentration of 18 ng.mL"1 [33]. It is important to note, however, that our findings do not exclude the possibility that thioredoxin mediated exchange of the D2 disulfide indirectly through redox control of another cell-surface protein. The observations presented herein indicate that manipulation of disulfide-bonds in secreted proteins may be an important means of regulating certain protein function. Two other examples of this means of controlling protein function are reduction of plasmin disulfide-bonds [9], which triggers formation of angiostatin (an inhibitor of blood vessel formation), and disulfide-bond reduction/oxidation in von Willebrand Factor [10], which regulates its haemostatic activity.

Acknowledgements We thank A. Holmgren and F. Clarke for providing the thioredoxins. This work was supported by the Australian Research Council and the National Health and Medical Research Council of Australia and the NSW Health Department.

References 1. R. Noiva, Enzymatic catalysis of disulfide formation. Protein Expr. Purif. 5 (1994) 1-13. 2. C.N. Pace, G.R. Grimsley, J.A. Thomson and B.J. Barnett, Conformational stability and activity of ribonucelase T1 with zero, one, and two intact disulfide-bonds. J. Biol. Chem. 263 (1988) 11820-11825. 3. R. Wetzel, Harnessing disulfide-bonds using protein engineering. Trends Biochem. Sci. 12 (1987) 478-482. 4. S.J. Weiner, P.A. Kollman, D.A. Case, U.C. Singh, C. Ohio, G. Alagona, S.Jr. Profeta and P. Weiner, A new force field for molecular mechanical simulation of nucleic acids and proteins. J. Am. Chem. Soc. 106 (1984)765-784. 5. B.A. Katz and A. Kossiakoff, The crystallographically determined structures of atypical strained disulfides engineered into subtilisin. J. Biol. Chem. 261 (1986) 15480-15485. 6. P.E. Pjura, M. Matsumura, J.A. Wozniak and B.W. Matthews, Structure of a thermostable disulfide-bridge mutant of phage T4 lysozyme shows that an engineered cross-link in a flexible region does not increase the rigidity of the folded protein. Biochemistry 29 (1990) 2592-2598. 7. J.A. Wells and D.B. Power, In vivo formation and stability of engineered disulfide-bonds in subtilisin. J. Biol. Chem. 261 (1986) 6564-6570. 8. J.M. Thornton, Disulphide bridges in globular proteins. J. Mol. Biol. 151 (1981) 261-287. 9. A.J. Lay, X.-M. Jiang, O. Kisker, E. Flynn, A. Underwood, R. Condron and P.J. Hogg, Phosphoglycerate kinase acts in tumour angtogenesis as a disulphide reductase. Nature 408 (2000) 869-873. 10. L. Xie, C.N. Chesterman and P.J. Hogg, Control of von Willebrand factor multimer size by thrombospondin1. J. Exp. Med. 193 (2001) 1341-1349. 11. R.J. Harris, S. M. Chamow, T.J. Gregory and M.W. Spellman, Characterization of a soluble form of human CD4. Eur. J. Biochem. 188 (1990) 291-300.

272

LJ. Matthias el at. / Disulfide Exchange in CD4

12. J. Wang, Y. Yan, T.P.J. Garrett, J. Liu, O.W. Rodgers, R.L Garlick, G.E. Tarr, Y. Husain, E.L Reinherz and S.C. Harrison, Atomic structure of a fragment of human CD4 containing two immunogtobuMrHike domains. Nature 348 (1990) 411-418. 13.S.R. Ryu, P.O. Kwong, A. Truneh, T.G. Porterm, J. Arthos, M. Rosenberg, X. Dai, N. Xuong, R. Axel, R.W. Sweet and W.A. Hendrickson, Crystal structure of an HIV-binding recombinant fragment of human CD4. Nature 348 (1990) 419-426. 14. R.L. Brady, E.J. Dodson, G. Lange, S.J. Davis, A.F. Williams and A.N. Barclay, Crystal structure of domains 3 and 4 of rat CD4: relation to the NH2-terminal domains. Science 260 (1993) 979-983. 15. N. Srinivasan, R. Sowdhamini, C. Ramakrishnan and P. Balaram, Conformation of disulfide bridges in proteins. Int. J. Peptide Protein Res. 36 (1990) 147-155. 16.1. Nakashima, M. Pu, A. Nishizaki, I. Rosila, L Ma, Y. Katano, K. Ohkusu, S.M.J. Rahman, K. Isobe, M. Hamaguchi and K. Saga, Redox mechanism as alternative to ligand binding for receptor activation delivering disregulated cellular signals. J. Immunol. 152 (1994) 1064-1071. 17. G.W. Lynch, A.J. Sloane, V. Rasco, A. Lai and A.L Cunningham, Direct evidence for native CD4 oligomers in lymphokJ and monocytoid cells. Eur. J. Immunol. 29 (1999) 2590-2602. 18. H.J.-P. Ryser, E.M. Levy, R. Mandel and G.J. DiSciulto, Inhibition of human immunodeficiency virus infection by agents that interfere with thiol-disulfide interchange upon virus-receptor interaction. Proc. Mail. Acad. Sci. USA 91 (1994) 45459-4563. 19. H.M. Ingalls, C.M. Goodloe-Holland and E.J. Luna, Junctional plasma membrane domains isolated from aggregating Dictyostelium discoideum amebae. Proc. Natl. Acad. Sci. USA 83 (1986) 4779-4783. 20. X.-M. Jiang, M. Fitzgerald, C.M. Grant and P.J. Hogg, Redox control of exofacial protein thiots/disutfides by protein disulfide isomerase. J. Biol. Chem. 274 (1999) 2416-2423. 21. K. Tonissen, J. Wells, I. Cock, A. Perkins, C. Orozco and F. Clarke, Site-directed mutagenesis of human thioredoxin. J. Biol. Chem. 268 (1993) 22485-22489. 22. K.C. Deen, J.S. McDougal, R. Inacker, G. Folena-Wasserman, J. Arthos, J. Rosenberg, P.J. Maddon, R. Axel, R.W. Sweet, A soluble form of CD4 (T4) protein inhibits AIDS virus infection. Nature 331 (1988) 8284.

23. J.A. Richardson, Protein anatomy. Advan. Protein. Chem. 34 (1981) 167-339. 24.1.L. Karie, R. Kishore, S. Raghothama and P. Balaram, Cyclic cystine peptides. Antiparattel p-sheet conformation for the 20-membered ring in Boc-Cys-Val-Aib-Ala-Leu-Cvs-NHMe. J. Am. Chem. See. 110 (1988)1958-1963. 25.1.L. Karie, J.L Flippen-Anderson, R. Kishore and P. Balaram, Cystine peptides. Int. J. Peptide Protein Res. 34(1989)37-41. 26. A.F. Williams, S.J. Davis, Q. He and A.N. Barclay, Structural diversity in domains of the immunoglobulin superfamiry. Cold Spring Harbor Symp. Quant. Biol. 65 (1989) 637-647. 27. L Kirszbaum, J.A. Shame, N. Goss, J. Lahnstein and I.D. Walker, The alpha-chain of murine CDS lacks an invariant Ig-like disulfide-bond but contains a unique intrachain loop instead. J. Immunol. 142 (1989) 39313936. 28. A. Holmgren, Thioredoxin. Ann. Rev. Biochem. 54 (1985) 237-271. 29. Y. Tagaya, Y. Maeda, A. Mitsui, N. Kondo. H. Matsui, J. Hamuro, N. Brown. K. Aral, T. Yokota. H. Wakasugi and J. Yodoi, ATL-derived factor (ADF), an IL-2 receptor/Tac inducer homologous to thioredoxin: possible involvement of a dithid-reduction in the IL-2 receptor induction. EMBO J. 8 (1989) 757-764. 30. H. Martin and M. Dean, Identification of a thioredoxin-related protein associated with plasma membranes. Biochem. Biophys. Res. Commun. 175 (1991) 123-128. 31. A. Rubartelli, A. Bajetto, G. Allavena, E. Wollman and R. Sitia, Secretion of thioredoxin by normal and neoplastic cells through a leadertess secretory pathway. J. Biol. Chem. 267 (1992) 24161-24164. 32. E.E. Woolman, A. Kahan and D. Fradelizi, Detection of membrane associated thioredoxin on human cell lines. Biochem. Biophys. Res. Commun. 230 (1997) 602-606. 33. A. Soderberg, B. Sahaf and A. Rosen, Thioredoxin reductase, a redox-active setenoprotein, is secreted by normal and neoplastic cells: presence in human plasma. Cancer Res. 60 (2000) 2281-2289.

Thiol Metabolism and Redox Regulation of Cellular Functions A. Pompella ct al. (Eds.) [OS Press. 2002

Redox Regulation in Protein Folding and Chaperone Function Peter CSERMELY, Gabor NARDAI and Csaba SOTI Department of Medical Chemistry, Semmelweis University, P.O. Box 260, H-1444 Budapest 8, Hungary 1. Introduction: Protein Folding, Heat Shock Proteins, Molecular Chaperones Chaperones are ubiquitous, highly conserved proteins, which utilize a cycle of ATP-driven conformational changes to refold their targets, and which probably played a major role in the molecular evolution of modern enzymes [1,2]. Environmental stress (a sudden change in the cellular environment, to which the cell is not prepared to respond, such as heat shock) leads to the expression of most chaperones, which therefore are called heat-shock, or stress proteins. Lacking a settled view about their action in the molecular level [3], chaperones are still best classified by their molecular weights (Table 1). Besides to promote the formation of the correct conformation of nascent or damaged proteins chaperones also assist in the formation of correct disulfide bridges offering the help protein disulfide isomerases (PDI-s) [4,5]. Higher levels of cellular organization also need a constant remodeling. Chaperones are obvious candidates to provide help in these processes. About twenty years ago based on high-voltage electron microscopy Keith A. Porter and co-workers suggested the existence of a cellular meshwork, called as "microtrabecular lattice" to organize cytoplasmic proteins and RNA-s [6]. Almost instantly a fierce debate arose considering the lattice as an artifact of the techniques used. However, as time passes more and more data provide indirect evidence for a high-order organization of the cytoplasm [7]. Chaperones are ideal candidates for being a major constituent of a cytoplasmic meshwork: they are highly abundant, form a loose and dynamic complex with all the elements of the cytoskeleton and each other, and also attach to a plethora of other proteins. Several lines of initial evidence shows that disruption of chaperone/protein complexes disturbs the organization of cytoplasmic traffic of several proteins, such as the steroid receptors, and accelerates cell lysis [8-10].

2. Redox Chaperones in the Endoplasmic Reticulum and in the Periplasm: Quality Control of Secreted Proteins Secreted proteins have to prepared for the oxidative millieu of the extracellular space. A rapid oxidation would result in the formation of numerous incorrect disulfide bridges, which would lock the protein in a distorted conformation. Therefore folding of secreted/plasma membrane proteins is most probably accompanied by their gradual oxidation in the endoplasmic reticulum (ER). This would imply the existence of a redox gradient along the secretory

274

P. Csermely etal./ Redox Regulation in Protein Folding and Chaperone Function

pathway. The first tools, such as the redox sensitive green fluorescent protein (GFP, which, in fact, has a yellow colour in this case, 11 ] to measure this putative gradient have already been established. Table 1. Major classes of molecular chaperones

a

Most important eukaryotic representatives8

Reviews

small heat shock proteins (e.g. Hsp27 ")

21

Hsp60

1,46

Hsp70 (HscTO c, Grp78)

1,46

Hsp90

8,47

Hsp104

48

Peptidyl-prolyl c/s/frans-isomerases

49

Protein disulfide isomerases

4,5

Co-chaperones (chaperones which help the function of other chaperones listed) were not included in this table, albeit almost all of these proteins also possess a "traditional" chaperone activity in their own right. Several chaperones of the endoplasmic reticulum (e.g. calreticulin, calnexin, etc.). which do not belong to any of the major chaperone families, as well as some heat shock proteins (e.g. ubiquitin), which do not possess chaperone activity were also not mentioned.

" The abbreviation "Hsp" and "Grp" refer to heat shock proteins, and glucose regulated proteins, chaperones induced by heat shock or glucose deprivation, respectively. Numbers refer to their molecular weight in kDa. 0

Hsc70 denotes the cognate (constitutivety expressed) 70 kDa heat shock protein homologue in the cytoplasm.

Another mechanism for the control of gradual oxidation is the reorganization of rapidly formed, incorrect disulfide bonds. This is performed by the numerous protein disulfide isomerases (PDI-s). These proteins have been discovered independently by the group of Christian Anfmsen [12], and by Pal Venetianer and Bruno Straub [13] in Hungary almost fourty years ago. In the meantime the existence of a large number of the family has been uncovered, such as Erp29, Erp57, Erp59, Erp72 and others. The exact substrate specificity of these enzymes is not known. However, Erp59 seems to be the most abundant member of the family and Erp72 acts mainly on glycosylated proteins. The formation of disulfide bridges requires the accessibility and correct positioning of SH groups. This is achieved by the "conventional" chaperone activity of protein disulfide isomerases [14] as well as by their cooperation with other chaperones, such as Grp78 in the ER [15]. The correct chaperone/target ratio is very important in the action of PDI-s. In case of a large excess of targets PDI acts as an "anti-chaperone" promoting inter- and not intramolecular disulfide bridge formation [16]. Therefore, in case of ER-overload or after the poisoning by reducing agents PDI-s may promote the formation of covalently linked aggregates instead of their usual role as redox-chaperones. Besides the reorganization of inccorect disulfide bridges protein disulfide isomerases also participate in the direct oxidation of secreted proteins. In these cases oxidized protein disulfide isomerases are reduced by the ER transmembrane proteins Erol-L-alpha and ErolL-beta requiring flavin adenine dinucleotide as a cofactor. The two Ero-s seem to act on

P. Csermely et al. /Redox Regulation in Protein Folding and Chapewne Function

different PDI-s, and the more active form, Ero 1 -L-beta is overregulated if the ER experiences an excess of unfolded proteins [17]. In contrast of yeast PDI, which is present predominantly in the oxidized state, most mammalian PDI-s are partially reduced. The small pool of oxidized PDI suggests that in mammalian cells oxidative equivalents are rapidly transferred to cargo proteins [17] or the redox gradient along the secretory pathway is more expressed than in yeast. Protein disulfide isomerases are not always ER-resident proteins. They are also secreted to the extracellular medium, where they continue their folding assistance, and prevent the formation of extracellular protein aggregates [18,19]. In the E. coli periplasm a slightly different mechanism controls the oxidation of proteins to that of the ER: oxidation is achieved by a separate arm of proteins involving DsbA, which is reoxidized by the inner membrane protein DsbB. The E. coli protein disulfide isomerases are DsbC and DsbG, which are reoxidized by the membrane protein DsbD [4,5]. Table 2. Cytoplasmic redox chaperones

Function

Chaperone small heat shock

increases reduced glutathione levels by

proteins (Hsp25,

increased glucose-6-phosphate

Hsp27)

dehydrogenase, glutathione reductase and

References 20,21

glutathione transferase activities Hsp32a

heme oxygenase-1, an important component of

22

oxidative stress-mediated cell injury Hsp33

oxidation-activated chaperone in yeast

23

Hsp70

has (probably indirect) anti-oxidant properties

50

cytochrome c

released from mitochondria in apoptosis,

27

chaperone function has been detected thioredoxin3

promotes cytoplasmic oxidation of selected

24

proteins ERV1/ALR3

promotes the cytoplasmic formation of disulfide

25

bridges MsrA and MsrBa

methionine sulfoxide reductase: regenerates

26

functional methionine 1

No direct chaperone activity has been demonstrated yet.

3. Redox Chaperones in the Cytoplasm: Another Defense against Oxidative Damage From the traditional chaperones (Table 1) the small heat shock proteins and Hsp70 act as cytoplasmic "antioxidants" (Table 2). Small heat shock proteins elevate reduced glutathione levels by promoting an increase in glucose-6-phosphate dehydrogenase activity and by a somewhat smaller activation of glutathione reductase and glutathione transferase [20,21]. Heme oxygenase a heat shock protein responsible for the production of the antioxidants

276

P. Csermely et al. /Redox Regulation in Protein Folding and Chaperone Function

biliverdin and bilirubin is an important component of cellular defense mechanisms against oxidative stress [22]. Yeast cells contain a rather unique chaperone, Hsp33, which is activated after oxidative stress [23]. SH-groups of some specific proteins (such as the glucocorticoid receptor) are maintained in the reduced state or just inversely: oxidized by thioredoxins in the cytoplasm [24]. In turn, these proteins are most probably oxidized by members of the ERV1/ALR protein family [25]. Oxidized methionines are reduced by special enzymes, the methionine sulfoxide reductases [26]. An interesting member of redox cytoplasmic chaperones is cytochrome c, which is only a "guest" in the cytoplasm during its apoptotic release from mitochondria and has been established as a chaperone a long time ago [27].

4. Redox Control of Chaperone Induction Oxidative stress leads to a massive induction of heat shock proteins. This is partly mediated by the oxidation-induced formation of damaged proteins [28], which occupy chaperonebinding sites, and liberate heat shock factor 1 (HSF1), the transcription factor responsible for Hsp induction [29]. However, a decrease in reduced glutathione levels (by oxidation or by the formation of S-nitroso-glutathione by NO) [30] may also lead to a direct activation of HSF1 [31]. Interestingly, a more reduced cellular environment also helps chaperone induction [32]. In agreement with this, heat shock itself leads to a rapid elevation of reduced glutathione [33]. On the other hand, millimolar concentration of a reducing agent impairs the activation of HSF1 [34].

5. Redox Regulation of Chaperone Function Cytoplasmic chaperones, such as small heat shock proteins, or Hsp90 usually loose their activity after the oxidation of their cysteine or methionine residues [35,36]. Both in Hsp70 and Hsp90 the oxidation-prone cysteine is in the close vicinity of an ATP binding site [36,37], which may explain the rapid loss of their chaperone activity after oxidation. Small heat shock proteins were recently established as ATP-binding chaperones, therefore the above explanation may have a more general implication than we previously thought. Chaperoneinactivation may also occur by S-nitrosylation after NO, or peroxonitrite addition. On the contrary to this general trend, the yeast cytoplasmic chaperone, Hsp33 is activated after oxidative stress [23]. Though its homologues have not been found in higher eukaryotes, a similar mechanism would be very logical to operate in other cells. Rapoport and co-workers [38] raised the interesting possibility that oxidation of PDI may trigger a release of its substrate, which would then travel further in the secretory chain or, as in the case of cholera toxin, would be a subject of a retrograde transport back to the cytoplasm.

6. Changes of Chaperones and Redox Function in Disease and Aging In several diseases, such as in endothelial dysfuncton, in diabetes, in Alzheimer and Parkinson diseases the redox homeostasis becomes severly damaged [39,40]. The amount of oxidized proteins increases, which requires a larger amount of chaperones to cope with the conformational damage and leads to chaperone induction. However, chronic stress exhausts the chaperone-induction signalling mechanisms, and damaged proteins begin to accumulate. Moreover, oxidized proteins are much poorer substrates, but highly effective inhibitors of the proteasome [41]. All these changes also occur during ageing with the

P. Csermely et al. / Redox Regulation in Protein Folding and Chaperone Function

777

concomitant decrease in the chaperone-induction capacity of the aging organism [42]. In contrast, enhanced antioxidant systems (such as the overexpression of Cu-Zn-superoxide dismutase) as well as an increased amount of heat shock proteins leads to longevity [43].

Fig. 1. Chaperones as major transmitters of changes in redox homeostasis to the life of the whole cell. Clockwise from bottom: (a) phenotypically buffered, silent mutations require the assistance of chaperones to rescue them from folding traps [44,45]. (b) Chaperones form low affinity and highly dynamic extensions of the cytoskeleton participating in cellular traffic and in the organisation of the cytoarchitecture [8-lo]. (c) Cytoplasmic chaperones of eukaryotic cells participate in the maintenance of the conformation of some, selected protein substrates. Most of these unstable proteins are parts of various signalling cascades [8,47]. (d) After changes in the redox homeostasis, chaperones become more and more occupied by damaged proteins. As a consequence of this: (a) silent mutations escape and contribute to the onset of polygenic diseases; (b) cell architecture becomes disorganised; (c) signalling is impaired. The verification of these - presently largely hypothetical - changes requires further experimentation.

7. Perspectives: Chaperones as Central Players in the Transmission of Redox Changes to the Life of the Cell Oxidative damage, together with other proteotoxic insults during the propagation of various diseases and ageing results in a change between the ratio between damaged proteins and available chaperone capacity. The chaperone-overload, which is a consequence of these events leads to rather unexpected changes. Recently one of the major cytoplasmic chaperones, Hsp90, was shown to act as posttranslational "silencer" of several genetical changes by assisting in an efficient repair of folding defects 1441. After a large stress transient chaperone-overload prevents the conformational repair of misfolded mutants. Therefore many, previously hidden genotypical changes appear in the phenotype resulting in a "boom" of genetical variations in the whole population. This may help the selection of a beneficial change, which, in turn, may help the adaptation of the population to the changing environmental conditions. Under stressful conditions most of the exposed mutations are disadvantageous, and tend to disappear from the population by natural selection. According to a recent hypothesis [45] the development of modern medical practice depressed natural selection by its groundbreaking achievements to reduce prenatal and infant mortality leading to a rise of phenotypically silent mutations in the genome. As a consequence we carry more and more

278

P. Csermely et al /Redox Regulation in Protein Folding and Chaperone Function

chaperone-buffered, silent mutations from generation to generation. The chance of the phenotypic manifestation of these mutations becomes especially large in aged subjects, where protein damage is abundant, and chaperone induction is impaired. The background of misfolded proteins increases and by competition prevents the chaperone-mediated buffering of silent mutations. Phenotypically exposed mutations contribute to a more abundant manifestation of multigene-diseases, such as atherosclerosis, autoimmune-type diseases, cancer, diabetes, hypertensive cardiovascular disease and several psychiatric illnesses (Alzheimer disease, schizophrenia, etc.). The "chaperone overload" hypothesis emphasises the need for efficient ways to enhance chaperone-capacity in ageing subjects, and calls for the identification and future "repair" of silent mutations [45]. After oxidative damage in the cytoplasm or "reductive damage" in the endoplasmic reticulum, the resulting chaperone overload changes the whole life of the cell. Besides the exposure of the previously hidden mutations signaling becomes impaired and the cellular architecture is disorganized (Figure 1). Chaperones may act as central players of the transmission of redox changes in the life of the cell. References 1. F-U. Hartl, Molecular chaperones in cellular protein folding. Nature 381 (1996) 571-580. 2. P. Csermely, Proteins, RNA-s, chaperones and enzyme evolution: a folding perspective. Trends in Biochem. Sci. 22 (1997) 147-149. 3. P. Csermely, The "chaperone-percolator" model: a possible molecular mechanism of Anfinsen-cage type chaperone action. BioEssays 21 (1999) 959-965. 4. S. Raina and D. Missiakas, Making and breaking disulfide bonds. Annu. Rev. Microbiol. 51 (1997) 179202.

5. A. Rietsch and J. Beckwith, The genetics of disulfide bond metabolism. Annu. Rev. Genetics 32 (1998) 163-184. 6. M. Schliwa, J. van Blerkom and K.R. Porter, Stabilization of the cytoplasmic ground substance in detergent-opened cells and a structural and biochemical analysis of its composition. Proc. Natl. Acad. Sci. USA 78 (1981) 4329-4333. 7. A.S. Verkman, Solute and macromolecule diffusion in cellular aqueous compartments. Trends in Biochem. Sci. 27 (2002) 27-33. 8. W.B. Pratt, A.M. Silverstein, and M.D. Galigniana, A model for the cytoplasmic trafficking of signalling proteins involving the hsp90-binding immunophilins and p50cdc37. Cellular Signaling 11 (1999) 839-851. 9. P. Csermely, A nonconventional role of molecular chaperones: involvement in the cytoarchitecture. News in Physiol. Sci. 15 (2001) 123-126. 10. A. Pato, K. Mihaly and P. Csermely, Chaperones and cytoarchitecture: geldanamycin induces an accelerated flux of cytoplasmic proteins from detergent-treated cells. Eur. J. Biochem. 268 (2001) 107a. 11. H. Ostergaard, A. Henriksen, F. G. Hansen and J. R. Winther, Shedding light on disutfide bond formation: engineering a redox switch in green fluorescent protein. EMBO J. 20 (2001) 5853-5862. 12. R.F. Goldberger, C.J. Epstein and C.B. Anfinsen, Acceleration of reactivation of reduced bovine pancreatic ribonuclease by a microsomal system from rat liver. J. Biol. Chem. 238 (1963) 628-635. 13. P. Venetianer and F. B. Straub, The enzymic reactivation of reduced ribonuclease. Biochim. Biophys. Acta 67(1963) 166-167. 14. Y. Yao, Y. Zhou and C. Wang, Both the isomerase and chaperone activities of protein disulfide isomerase are required for the activation of reduced and denatured acidic phospholipase A2. EMBO J. 16 (1997) 651658.

15. M. Mayer, U. Kies, R. Kammermeyer and J. Buchner, BiP and PDI cooperate in the oxidative folding of antibodies in vitro. J. Biol. Chem. 275 (2000) 29421-29425. 16. A. Puig and H.F. Gilbert, Protein disulfide isomerase exhibits chaperone and anti-chaperone activity in the oxidative refolding of lysozyme. J. Biol. Chem. 269 (1994) 7764-7771. 17. A. Mezghrani, A. Fassio, A. Benham, T. Simmen, I. Braakman and R. Sitia, Manipulation of oxidative protein folding and PDI redox state in mammalian cells. EMBO J. 20 (2001) 6288-6296. 18. H. Takemoto, T. Yoshimori, A. Yamamoto, Y. Miyata, I. Yahara, K. Inoue and Y. Tashiro, Heavy chain binding protein (BiP/grp78) and endoplasmin are exported from the endoplasmic reticulum in rat exocrine pancreatic cells, similar to protein disulfide-isomerase. Arch. Biochem. Biophys. 296 (1992) 129-136.

P. Csermely et al, /Redox Regulation in Protein Folding and Chaperone Function

279

19. P. Delom, B. Mallet, P. Carayon and P.J. Lejeune, Role of extracellular molecular chaperones in the folding of oxidized proteins. Refolding of colloidal thyroglobulin by protein disulfide isomerase and immunoglobulin heavy chain-binding protein.. J. Biol. Chem. 276 (2001) 21337-21342. 20. A. P. Arrigo, Small stress proteins: Chaperones that act as regulators of intracellular redox state and programmed cell death. Biol. Chem. 379 (1998) 19-26. 21.X. Preville, F. Salvemini, S. Giraud, S. Chaufour, C. Paul, G. Stepien, M.V. Ursini and A.P. Arrigo, Mammalian small stress proteins protect against oxidative stress through their ability to increase glucose6-phosphate dehydrogenase activity and by maintaining optimal cellular detoxifying machinery. Exp. Cell Res. 247(1999)61-78. 22. K.K. Elbirt and H.L. Bonkovsky, Heme oxygenase: Recent advances in understanding its regulation and role. Proc. Assoc. Am. Physicians 111 (1999) 438-447. 23. U. Jakob, W. Muse, M. Eser and J.C. Bardwell, A chaperone activity with a redox switch. Cell 96 (1999) 341-352. 24. K. Okamoto, H. Tanaka, H. Ogawa, Y. Makino, H. Eguchi, S. Hayashi, N. Yoshikawa, L. Poellinger, K. Umesono and I. Makino, Redox dependent regulation of nuclear import of the glucocorticoid receptor. J. Biol. Chem. 274 (1999) 10363-10371. 25.T.G. Senkevich, C.L. White, E.V. Koonin and B. Moss, A viral member of the ERV1/ALR protein family participates in a cytoplasmic pathway of disulfide bond formation. Proc. Natl. Acad. Sci. USA 97 (2000) 12068-12073. 26. R. Grimaud, B. Ezraty, J.K. Mitchell, D. Lafitte, C. Briand, P.J. Derrick and F. Barras, Repair of oxidized proteins. Identification of a new methionine sulfoxide reductase. J. Biol. Chem. 276 (2001) 48915-48920. 27. C.K. Ramakrishna Kurup, T.S. Raman, J. Jayaraman and T. Ramasarma, Oxidative reactivation of reduced ribonuclease. Biochim. Biophys. Acta 93 (1964) 681-683. 28. AT. McDuffee, G. Senisterra, S. Huntley, J.R. Lepock, K.R. Sekhar, M.J. Meredith, M.J. Borrelli, J.D. Morrow and M.L. Freeman, Proteins containin non-native disulfide bonds generated by oxidative stress can act as signals for the induction of the heat shock response. J. Cell. Physiol. 171 (1997) 143-151. 29. R.I. Morimoto, Regulation of the heat shock transcriptional response: cross talk between a family of heat shock factors, molecular chaperones, and negative regulators. Genes Dev. 12 (1999) 3788-3796. 30. Y.M. Kim, M.E. de Vera, S.C. Watkins and T.R. Billiar, Nitric oxide protects cultured rat hepatocytes from tumor necrosis factor-alpha-induced apoptosis by inducing heat shock protein 70 expression. J. Biol. Chem. 272(1997) 1402-1411. 31. M.R. Jacquier-Sarlin and B.S. Polla, Dual regulation of heat-shock transcription factor (HSF) activation and DNA-binding activity by H2O2: role of thioredoxin. Biochem. J. 318 (1996) 187-193. 32. K. Kato, H. Ito and K. Okamoto, Modulation of the arsenite-induced expression of stress proteins by reducing agents. Cell Stress Chaperones 2 (1997) 199-209. 33. J.B. Mitchell and A. Russo, Thiols, thiol depletion, and thermosensitivity. Radial. Res. 95 (1983) 471-485. 34. I.E. Huang, H. Zhang, S.W. Bae and A.Y. Liu, Thiol reducing reagents inhibit the heat shock response. J. Biol. Chem. 269 (1994) 30718-30725. 35. N. Gustavsson, B.P. Kokke, B. Anzelius, W.C. Boelens and C. Sundby, Substitution of coserved methionines by leucines in chloroplast small heat shock protein results in loss of redox response but retained chaperone-like activity. Protein Sci. 10 (2001) 1785-1793. 36. G. Nardai, B. Sass, J. Eber, Gy. Orosz and P. Csermely, Reactive cysteines of the 90 kDa heat shock protein, Hsp90. Arch. Biochem. Biophys. 384 (2000) 59-67. 37. Cs. Soli, A. Racz and P. Csermely, A nucleotide-dependent molecular switch controls ATP binding at the C-terminal domain of Hsp90: N-terminal nucleotide binding unmasks a C-terminal binding pocket. J. Biol. Chem. 277 (2002) 7066-7075 38. A. Tsai, C. Rodighiero, W.I. Lencer and T.A. Rapoport, Protein disulfide isomerase acts as a redoxdependent chaperone to unfold cholera toxin. Cell 104 (2001) 937-948. 39. G. Nardai, L. Braun, M. Csala, V. Mile, P. Csermely, A. Benedetti, J. Mandl and G. Banhegyi, Protein disulfide isomerase and protein thiol dependent dehydroascorbate reduction and ascorbate accumulation in the lumen of the endoplasmic reticulum. J. Biol. Chem. 276 (2001) 8825-8828. 40. L.M. Sayre, M.A. Smith and G. Perry, Chemistry and biochemistry of oxidative stress in neurodegenerative disease. Curr. Med. Chem. 8 (2001) 721-738. 41. B. Friguet, E.R. Stadtman and L.I. Szweda, Modification of glucose-6-phosphate dehydrogenase by 4hydroxy-2-nonenal. Formation of cross-linked protein that inhibits the multicatalytic protease. J. Biol. Chem. 269 (1994) 21639-21643. 42. D.M. Hall, G.L Saltier, C.A. Sattler, H.J. Zhang, L.W. Oberley, H.C. Pilot and K.C. Kregel, Aging lowers steady-state antioxidant enzyme and stress protein expression in primary hepatocyles. J. Gerontol. 56A (2001)8259-8267. 43. R. Arking, V. Burde, K. Graves, R. Hari, E. Feldman, A. Zeevi, S. Soliman, A. Saraiya, S. Buck, J. Vettraino, K. Sathrasala, N. Wehr and R.L. Levine, Forward and reverse selection for longevity in Drosophila is characterized by alteration of antioxidant gene expression and oxidalive damage patterns. Exp. Gerontol. 35 (2000) 167-185.

280

P. Csermely et al I Redox Regulation in Protein Folding and Chapemne Function

44. S.L Rutherford and S. Lindquist, Hsp90 as a capacitor for morphological evolution. Nature 396 (1998) 336-342. 45. P. Csermely, Chaperone-overtoad as a possible contributor to "civilization diseases": atherosclerosis, cancer, diabetes. Trends in Genetics 17 (2001) 701-704. 46. A. Bukau and A.L. Norwich, The Hsp70 and Hsp60 chaperone machines. Cell 92 (1998) 351-366. 47. P. Csermely, T. Schnaider, Cs. Soti, Z. Prohaszka and G. Nardai, The 9CTkDa molecular chaperone family: structure, function and clinical applications. A comprehensive review. Pharmacol. Therapeut. 79 (1998) 129-168. 48. J. Porankiewicz, J. Wang and A.K. Clarke, New insights into the ATP-dependent Clp protease: Escherichia coli and beyond. Mol. Microbiol. 32 (1999) 449-458. 49. S.F. Gothel and M.A. Marahiel, Peptidyl-prolyl cis-trans isomerases, a superfamily of ubiquitous folding catalysts. Cell. Mol. Life Sci. 55 (1999) 423-436. 50. Y.M. Kim, M.E. de Vera, S.C. Watkins and T.R. Billiar, Nitric oxide protects cultured rat hepatocytes from tumor necrosis factor-alpha-induced apoptosis by inducing heat shock protein 70 expression. J. Biol. Chem. 272 (1997) 1402-1411.

Thiol Metabolism and Redox Regulation of Cellular Functions A Pornpella et al. (Eds.) IOS Press, 2002

281

Reduction of the Endoplasmic Reticulum Accompanies the Oxidative Damage of Diabetes Mellitus Gabor NARDAI, Tamas KORCSMAROS and Peter CSERMELY Department of Medical Chemistry, Semmelweis University - Budapest, Hungary

1. Thiol Metabolism in the Endoplasmic Reticulum The average redox potential of the endoplasmic reticulum (ER) is about -160 mV, but theoretical calculations and some experimental results suggest that redox potential gradients and redox potential inhomogenities are typical of the subcompartment. The ER redox potential was thought to be mantained mostly by the glutathione/glutathione-disulfide redox buffer (GSH:GSSG = 1 trough 3:1), and can be described by the thiol/disulfide ratio [1]. But there are many other systems, which are able to influence the thiol metabolism. The in vitro mechanism to alter the redox state includes sulfhydryl oxidase, a NADPH-dependent oxigenase and the vitamin-K redox cycle [2]. The direct role of hem, ubiquinone, Fe-S clusters and molecular oxygene was excluded recently in yeast models [3]. The possible involvement of flavin adenine dinucleotide (FAD) in the electron transport was also documented on yeast, where the addition of FAD accelerated the disulfide bridge formation by the ER-resident enzyme, EROlp [3]. The yeast lumenal ER protein, ERV2 was also described as a flavoenzyme, and is able to accelerate O2-dependent disulfide bridge formation [4]. Besides flavin adenine dinucleotide, increasing number of evidence supports the involvement of an other, well known redox system on ER redox state. Ascorbate/dehydroascorbate concentration is in milimolar range in the ER lumen, and asorbate is a very important cofactor of the enzymes catalyzing prolyl- and lysylhydroxylation. The characteristics of their transport was well described by G. Banhegyi and co-workers [5]. Cytoplasmic ascorbic acid is first oxidized and dehydroascorbic acid is transported to the lumen by facilitated diffusion. Inside, the increasing concentration of oxidized form can help to mantain the transitional redox state, and take part in the GSH and protein thiol oxidation [6]. Moreover, protein disulfide isomerase itself has dehydroascorbate reductase activity [7]. The membrane-bound antioxidant agent, tocopherol was also mentioned as a possible contributor of the electron transport by ascorbic acid [8]. The importance of the glutathione/glutathione-disulfide redox buffer was recently described. The estimated redox potential of the ER (-160 mV) correlates with the current GSH/GSSG ratio, and its total concentration (1 to 2 mM) is high enough to affect redox environment and protein redox states [9]. But the processes setting the balance between GSH and GSSG have not yet been clearly identified. Glutathione synthetase, responsible for the de novo GSH generation, is located only in the cytoplasm, so GSH must enter to the ER lumen through transporters. A much faster GSSG transport was hypothesized to sustain the oxidative environment, but recent data are quite controversial

G. Nardai el al. /Reduction of the Endoplasmic Reticulum

282

on this topic [5]. Another possible way for the increase of GSSG concentration is described by some publications involving specific enzymatic GSSG generation by intraluminal redox enzymes [10]. The next chapter of glutathione research, its role on the protein folding process, is also under reevaluation today. The increasing importance of GSH is underlined as a counterbalance for the EROlp-mediated oxidation in yeast. According to this model, the oxidizing equivalents (whose precise nature is still unidentified) coming from the cytosol, and transmitted through the ER membrane by the EROlp protein, oxidize protein disulfide isomerase and consequently the secreted proteins. To avoid hyperoxidizing conditions a part of these oxidized proteins (and perhaps EROlp itself) are reduced by GSH and this process takes part in the maintenance of the relatively low ER GSH/GSSG ratio [11]. The final step of ER glutathione metabolism is the secretion of GSH, which occurs by the vesicular transport system. The concentration of the glutathione in the vesicles targeting secreted and membrane surface proteins to the cell membrane is about 1 mM [12].

Oxidant

C

Secreted proteins V g

GSH

^GSH ^GSSG Secreted proteins

ER lumen

Fig. 1. Redox balance in the ER [10]

2. Disulfide bond formation in the ER The cytoplasmic redox potential (-230 mV) and the high GSH/GSSG ratio (30 through 100:1) is not favourable to the formation and existence of disulfide bridges. Therefore, only 0.1% of the total protein thiol groups is bound in intramolecular or mixed disulfides permanently. Transitional disulfide bridges can be observed in case of many enzymes. This function is connected to a thiol-disulfide catalytic cycle [13,14]. Therefore, folding of the secreted and integral membrane proteins, which usually have many disulfide bridges to stabilize their tertiary and quaternary structure, must occur in the more oxidizing environment of the ER. The newly synthetized polypeptide chain labeled by the N-

G. Nardai et al. /Reduction of the Endoplasmic Reticulum

283

terminal signal peptide targeting it to the ER, enters to the lumen, where molecular chaperones and folding catalysts help to get its final structure. Molecular chaperones, like BiP and Grp94 trap the unfolded protein and stabilize it against aggregation, incorrect folding, but don't participate in the disulfide bond synthesis directly [15,16]. The formation of native disulfide bridges may occur without enzymatic catalysis under highly diluted and mild oxidative conditions as Christian Anfinsen and co-workers demonstrated almost 40 years ago [17], but the redox folding assistance, such as protein disulfide isomerases can accelerate and make the thiol-disulfide exchange more efficient in the crowded and dinamically changing environment of the ER. These proteins have some common caracteristics. They are intraluminal, mostly soluble enzymes containing the ER retention signal, and at least one thioredoxin-domain and a special amino acid sequence (CXXC) in their active site. The major steps of the enzymatic catalysis are the formation of a transitional enzyme-substrate complex, connected by intermolecular disulfide bridges and then the rearrangement of these bounds. Protein disulfide isomerases can create, cleave and isomerize the disulfide bonds depending on the state of the substrate and the redox eqilibrium[18,19].

Disulfide bond formation

Substrate-SH SH

+

S-PDI \S

*> Substrate-S-S-PDI S"H SH

> Substrate-S + SH-PDI V SH

+

S-PDI \S

SH ^Substrate-S-S-PDI SH SH

SH *- Substrate-S + SH-PDI V SH

Isomerisation

S Substrate-S SH

Fig. 2. The mechanism of thiol-disulfide exchange

The best known member of the PDI family is the 58 kDa protein disulfide isomerase, which is usually called as Erp58, or simply as protein disulfide isomerase (PDI). This PDI contains two thioredoxin domains, and its local concentration in the ER is about 200 to 300 uM [19,20]. It has both "molecular chaperone" and "antichaperone" activities on different substrates. As a chaperone it prevents protein aggregation. Just conversely, as an antichaperone PDI mediates aggregate formation, when the amount of unfolded or aggregation-prone proteins is far greater than that of ER chaperones [21]. PDI seems to be a redox-regulated chaperone: binding of the cholera-toxin substrate to PDI is stronger, when the protein is reduced and cholera toxin dissociates under oxidizing conditions [22]. PDI is also a constant part of various enzyme complexes assisiting in quite different processes (e.g. prolyl-4-hydroxilase, tryacylglycerol transfer protein, etc.) and can bind the hormone, estrogen [23]. But the most important PDI function is to form native disulfide bridges on newly syntethized and reversibly denatured proteins. This is a circular process and the reduced PDI is reoxidized by the ER membrane protein EROlp [24]. EROpl was discovered a couple of years ago in yeast, but later mammalian isoforms were also

284

G. Nardai et ai /Reduction of the Endoplasmic Reticulum

identified. EROpl contains a CXXCXXC motif in its active site, and is able to oxidize both GSH and PDI. PDI was found to bind directly to the EROlp. Neither structural, nor functional interaction between the EROlp and other members of the protein disulfide isomerase family, Erp72, Erp57 were detectable so far. In mammals, the two human EROlp isoforms (a and (5) were cloned, and only the expression of p* form is induced by the unfolded protein response. This is a flavoprotein, which is able to accept electrons coming from the secretory and other substrate proteins through protein disulfide isomerase and donate them to unknown participants [10]. ERp72 is a protein disulfide isomerase containing three thioredoxin domains. By its active redox and disulfide isomerase activity can replace PDI in PDI-deficient cells. In vivo it binds calcium, and the in vitro binding of different peptides could be also demonstrated together with some chaperone-like activities [20,25]. Erp57 is a protein disulfide isomerase specialized to glycoproteins. It has two thioredoxin domains, and forms complexes by calnexin and calreticulin the two ER lectinlike chaperones. The isomerase activity of the calreticulin/ or calnexin/Erp57 complex is higher than that of Erp57 alone. Erp57 interacts specifically with N-glycosylated polypeptides and its function is strongly determined by the lectin-like chaperones and the glycosylation-state of the substrate [26,27]. ERp44 is the newest member of the thioredoxin domain-containing ER luminal proteins, but the primary structure of its active site (CRFS) is different from that of other PDI-s in the ER. Erp44 helps immunoglobulin folding, binds covalently to EROlp and can influence its redox state [28]. There are many ER-resident proteins (Erp28, P5, Erp55) associated to these folding catalysts, which lack the thioredoxin domain, and they do not likely play a direct role in redox folding. However, they are suggested as possible co-chaperones of the redox folding process [29,30]. According to our current knowledge two different ways of disulfide bond formation are present in the ER. In the first, the electron transport occurs from the reduced substrate through protein disulfide isomerase to the EROlp. The EROlp flavin adenine dinucleotide group gives it to unidentified participants, which can integrate the thiol metabolism to the general cellular redox metabolic pathways. The second way (performed mostly by ERp72, and Erp57) seems to be uncoupled from EROlp, and requires other, probable small molecular mediators [31].

3. Disturbances of Redox Protein Folding As we discussed above, the redox regulation of the ER is closely connected to redox protein folding. Therefore, it is not surprising that any major disturbances of the ER milieu, especially that of the thiol metabolism can lead to inefficient protein folding, and the consequent accumulation of non-native, aggregation-prone polypeptides. These changes induce ER stress, and provoke the unfolded protein response [32]. Unfolded protein response (UPR) was first characterized in yeast, but we have an increasing number of data about the mammalian pathway [33]. In this system the increased concentration of the unfolded proteins indicates the danger, and the cell, by a complex signal transduction mechanism, accelerates the expression of different ER proteins, which are necessary to survive and handle non-native proteins. The sensors of the UPR are ER

G. Nardai ct al. /Reduction of the Endoplasmic Reticulum

285

transmembrane proteins, which are able to bind unfolded proteins with their luminal domain. In case of ER stress, ER chaperones become overloaded, and free unfolded proteins emerge occupying the sensor binding sites. PERK is a stress-induced protein kinase, it can phosphorylate eiF-2a, and thus it prevents any further protein synthesis under ER stress. However, the synthesis of UPR induced proteins must be proceeded via a different mechanism and is not inhibited by PERK. ATF6 has a cytosolic transactivation domain, which is cleaved from the ER upon stress, goes to the nucleus and helps the transcription of genes containing ER stress response element (ERSE) [34]. Irel proteins have a cytoplasmic ribonuclease activity. In ER stress they undergo oligomerization and phosphorylation, and are suspected to splice the XBP-1 transcription factor mRNA resulting in a more efficient XBP-1 translation and concomitant induction of ERSE-regulated genes [35]. UPR induces molecular chaperones, members of the redox folding pathway, glycoprotein folding catalysts, enzymes of lipid metabolism, participants of the vesicle transport, protein translocation and ER-associated protein degradation [36]. The ER overload response (EOR) is induced, when misfolded or normal proteins (because of the accelerated expression of viral proteins) accumulate in the ER lumen, fulfill the space, and disturb the normal function of the ER [37]. In EOR, besides the induction of chaperones, a special pathway is also activated, and generates an immune response against the infected of disorganized cell. ER overload induces a Ca++-efflux to the cytosol, followed by reactive oxigen species generation. The increased concentration of ROS will activate the NFicB transcription factor, which is responsible for the induction of several inflammatory genes, such as interferon, interleukines and other cytokines. Besides these changes the expression of the other proteins participating in antigen presentation is also increased [36].

Table 1. Agents that activate UPR or EOR UPR

2-Deoxyglucose Tunicamycin Brefeldin A Buthylhydroperoxide Dithiothreitol Heavy metal ions Calcium ionophores Castanospermine Cycloheximide TNFalpha Overexpression of proteins

EOR

286

G. Nardai el al. /Reduction of the Endoplasmic Reticulum

4. Changes of the ER Redox Status in Diabetes Mellitus and in Other Pathological States Several in vitro studies prove that the changes of the thiol/disulfide ratio, inhibition of the electron transport disrupt the redox folding process. On the contrary, many other factors, which can inhibit the proper protein folding by different mechanisms, will alter the redox balance. Thus, the redox folding process and thiol metabolism are closely connected in the ER. Although the interrelations of the different processes were recognized, the possible folding consequences of the pathological states, which are suspected to affect the redox state of the cellular compartments have been largely unnoticed. It was previously shown that the presence of reducing agents (DTT, 0mercaptoethanol) or strong oxidants (butylhydroperoxide, metal ions) in the cell culture medium blocks disulfide bridge formation [38]. Oxidative/reductive changes of the cellular environment were observed in many diseases as the cause or the consequence of the pathological state. The redox potential is hardly mantained upon hypoxic conditions. Acute or chronic hypoxia is general in many common diseases (ischemic injury of the tissues - myocardial infarction, stroke; tumor growth; pulmonary diseases; poisoning). Here, not only the low O2 concentration, but the ATP depletion and the reperfusion-induced generation of reactive oxygen species (ROS) all disturb the normal redox balance [39]. The oxidative stress is characteristic to inflammations also, where ROS are produced by the phagocytes. Drugs can also influence thiol metabolism. NO-donor chemicals (nitrogycerol, nitroprusside sodium, etc) require free protein thiols to act on the vascular smooth muscles. Besides their direct effect we should keep in mind that many drugs are metabolised by the microsomal oxigenases, the P450 proteins, and that these redox enzymes might be in contact with other redox systems than the thiol/disulfide pathway [40]. The diabetes mellitus is described as a complex metabolic disease characterized by the absolute or relative shortage of insulin. One of the consequences of the metabolic disorganization is the increased generation of ROS [41,42]. The oxidative stress is initially prevented by the different antioxidative defense systems, but later on these mechanisms become exhausted, and oxidative damage develops. The elevated concentration of the oxidative stressors is the result of the disorganized function of the mitochondria, glucose autooxidation, and the free radicals generated by the non-enzymatically glycated proteins. These changes are typical of the extracellular space, but the signs of the oxidative stress are also detectable in the cytosol [39]. The thiol metabolism might be affected by other factors too. There are several evidences showing that the intracellular level of ascorbic acid, tocopherol and FAD is lower in both diabetic animal models and diabetic patients [43]. These changes are caused partly because of the accelerated cofactor consumption, and partly by the deficient cofactor-transport. Both the plasmamembrane transporter of ascorbic acid and dehydroascorbic acid and their microsomal uptake are blocked by high glucose concentration [5,44] The activity of some FAD-containing enzymes is significantly lower in experimental diabetes. These small molecules are all suggested to take part in the thiol metabolism of the ER, and more experiments required to decide if they are responsible for the suspected disturbances of the redox state and protein folding .

G. Nardai et al. /Reduction of the Endoplasmic Reticulum

287

In our recent work examining streptozotocin induced diabetic rats, we found that in spite of the oxidative changes of the extracellular space, the redox environment of liver microsomal vesicles was shifted to more reducing state [45]. Diabetic microsomal redox status was characterized by an increased total disulfide content and by and increased protein-thiol:disulfide ratio. The dehydroascorbate reductase activity was also higher in diabetes [45]. The cytoplasmic parameters remained to be unchanged. A similar redox shift was also observed in some ER proteins involved in the disulfide bond formation. Importantly, a significant portion of the protein disulfide isomerase and ERp72 were found to be in reduced form (data not shown). These changes were detected by the verification of the mobility shift resulted by the covalent modifications of SH-groups on the reduced proteins [46]. In diabetic samples we found more PDI-containing aggregates, which were only partially sensitive to reducing agents (data not shown).

5. Discussion and Perspectives Redox changes in diabetes alter the structure and perhaps the function of the ER redox protein folding-machinery in liver. That might be a possible explanation of the decreased hepatic protein secretion observed in some studies [47], and can also contribute to the accelerated protein turnover detected upon oxidative stress conditions [48]. The exact consequences of the more reducing environment are not clear yet. However, the regulatory role of the redox state is well known in many processes. The chaperoning activity of the PDI was found to be redox sensitive, the accumulation of the reduced form can lead the formation of more noncovalently bound complexes and a slowdown in redox-chaperoning [22]. The reducing conditions can also decrease protein stability and increase ER protein degradation. These redox changes in diabetes may influence the transport, presence and redox function of extracellular PDI on the plasma membrane [49]. We do not know, how typical are the changes of the ER metabolism we observed in diabetes mellitus. Are these changes common consequences of the oxidative stress or of any other aspecific events? More experiments and the use of different models is necessary to answer this question, and to understand the mechanisms involved in the pathology of the ER thiol metabolism, and to their evaluation as possible therapeutic targets in various diseases. References [1] [2] [3] [4] [5] [6] [7]

C. Hwang, A.J. Sinskey and H.F. Lodish. Oxidized redox state of glutathione in the endoplasmic reticulum. Scienc.e 257 (1992) 1496-1502. D.M. Ziegter and L.L. Poulsen. TiBS. 2 (1977) 79. B.P. Tu, S.C. Ho-Schleyer, K.J. Travers and J.S Weissman. Biochemical basis of oxidative protein folding in the endoplasmic reticulum. Science. 290 (2000) 1571-74. C.S. Sevier, J.W. Couzzo, A. Vala, F. Aslund and C.A. Kaiser. A flavoprotein oxidase defines a new endoplasmic reticulum pathway for biosynthetic disulfide bond formation. Nat Ceil Biol. 3 (2001) 874-882. G. Banhegyi, P. Marcolongo, F. Puskas, R. Fulceri, J. Mandl and A. Benedetti. Dehydroascorbate and ascorbate transport in rat liver microsomal vesicles. J Biol Chem. 273 (1998) 2758-2762. M. Csala, V. Mile, A. Benedetti, J. Mandl and G. Banhegyi. Ascorbate oxidation is a preequisite for its transport into rat liver microsomal vesicles. Biochem J. 349 (2000) 413-15. W.W. Wells, D. Peng Xu, Y. Tang and P.A. Rocque. Mammalian thiotransferase (glutaredoxin) and protein disulfide isomerase have dehydroascorbate reductase activity. J Biol Chem. 265 (1990) 15361-

64.

288

[8]

[9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] [33] [34] [35] [36] [37]

G. Nardai et al. /Reduction of the Endoplasmic Reticulum

M. Csala, A. Szarka, £. Margittai, V. Mite, T. Kardon, L Braun, J. Mand and G. Banhegyi. Role of vitamin E in ascorbate-dependent protein thiol oxidation in rat liver endoplasmic reticulum. Arch Biochem Biophys. 388 (2001) 55-59. S. Raina and D. Missiakas. Making and breaking disulfide bonds. Annu Rev Microbiol. 51 (1997) 179-202. A.R. Frand, J.W. Couzo and C.A. Kaiser. Pathways for protein disulfide bond formation. Trend Cel Biol. 10(2000)203-210. J.W. Couzzo, C.A. Kaiser. Competition between glutathione and protein thiols for disulphide-bond formation. Nat Cell Biol. 1 (1999) 130-135. Cysteine and glutathione secretion in response to protein disulfide bond formation in the ER. Science. 277(1997)1681^84. V.P. Pigiet and B.J. Schuster. Thioredoxin-catalyzed refolding of disulfide containing proteins. Proc NatJ Acad Sci. 83 (1986) 7643-47. E.S.J. Amerand A. Holmgren. Physiological functions thioredoxin and thioredoxin reductase. Eur J Biochem. 267 (2000) 6102-09. M. Mayer, U. Kies, R. Kammermeier and J. Buchner. BiP and PDI cooperate hi the oxidative folding of antibodies hi vitro. J Bio Chem. 275 (2000) 29421-25. Amy S. Lee. The glucose-regulated proteins: stress induction and clinical application. TIBS. 26 (2001) 504-510. R.F. Goldbeger, C.J. Epstein and C.B. Anfinsen. Acceleration of reactivation of reduced bovine pancreatic ribonuclease by a microsomal system from rat liver. J Biol Chem. 238 (1963) 628-635. M. Narayan, E. Welker, W.J. Wedemeyer and H.A. Scheraga. Oxidative folding of proteins. Ace Chem Res. 33 (2000) 805-12. R.B. Freedman, T.R. Hirst and M.F. Tuite. protein disulfide isomerase: building bridges hi protein folding. TiBS. 19(1994)331-36. D.M. Ferrari and H.O. Sdling. The protein disulphide-isomerase family: unravelling a string of folds. Biochem J. 339 (1999) 1-10. A. Puig and H.F Gilbert. Protein disulfide isomerase exhibits chaperone and anti-chaperone activity in the oxidative refolding of lysozyme. J Bid Chem. 269 (1994) 7764-71. B. Tsai, C. Rodighiero, W.I. Lencer and T.A. Rapoport. Protein disulfide isomerase acts as a redoxdependent chaperone to ubfold cholera toxin. Cell. 104 (2001) 937-48. T.P. Primm and H.F. Gilbert. Hormone binding by protein disulfide isomerase, a high capacity hormone reservoir of the endoplasmic reticulum. J Biol Chem. 276 (2001) 281-86. A.R. Frand and C.A. Kaiser. The ERO1 gene of yeast is required for oxidation of protein dithiols in the endoplasmic reticulum. Mol Cell. 1 (1998) 161-170. R.A. Mazarella, M Srinivasan, S.M. Haugejorden and M. Green. Erp72, an abundant luminai endoplasmic reticulum protein, contains three copies of the active site sequences of protein disulfide isomerase. J Biol Chem. 265 (1990) 1094-1101. S. High, F.J.L. Lecomte, S.J. Russel, B.M. Abell and J.D. Oliver. Glycoprotein folding hi the endoplasmic reticulum: a tale of three chaperones. FEBS Lett. 476 (2000) 38-41. M. Molinari and A. Helenius. Glycoproteins form mixed disulfides with oxidoreductases during folding in living cells. Nature. 402 (1999) 90-93. T. Anelli, M. Alessio, A. Mezghrani, T. Simmen, F. Talamo, A. Bachi and R. Sitia. Erp44, a novel endoplasmic reticulum folding assistant of the thioredoxin family. EMBO J. 21 (2002) 835-44. K. Weis. G. Griffiths and A.I. Lamond. The endoplasmic reticulum calcium-binding protein of 55 kDa is a novel EF-hand protein retained in the ER by acarboxyl-terminal His-Asp-Glu-Leu motif. J Biol Chem. 269 (1994) 19142-150. D.M. Ferrari, P. Nguyen Van, H.D. Kratzin and H.D. Sdling. Erp28, a human endoplasmic-reticulumlumenal protein, is a member of the protein disulfide isomerase family, but lacks a CXXC thtoredoxin-box motif. Eur J Biochem. 255 (1998) 570-79. A. Mezghrani, A. Fassio, A. Benham, T. Simmen, I. Braakman and R. Sitia. Manipulation of oxidative protein folding and PDI redox state in mammalian cells. EMBO J. 20 (2001) 6288-96. Randal J. Kaufman. Stress signaling from the lumen of the endoplasmic reticulum: coordination of gene transcriptional and translational controls. Genes Dev. 13 (1999) 1211-1233. Y. Ma and L.M. Hendershot. The unfolding tate of the unfolded protein response. CeH. 107 (2001) 827-30. Yoshida, K. Haze, H. Yanagi, T. Yura and K. Mori. Identification of the cis-acting endoplasmic reticulum stress response element responsible for transcriptional induction of mammalian glucose-regulated proteins. Involvement of basic leucine zipper transcription factors. J Biol Chem. 273 (1998) 33741-79. Yoshida, T. Matsui, A. Yamamoto, T. Okada and K. Mori. XBP1 mRNA is induced by ATF6 and spliced by IRE1 in response to ER stress to produce a highly active transcription factor. Cell. 107 (2001) 881-91. K.J. Travers, C.K. Patil, L. Wodicka, D.J. Lockhart, J.S. Weisman and P. Walter. Functional and genomic analyses reveal an essential coordination between the unfolded protein response and ERassociated degradation. Cett 101 (2000) 249-258. Heike L Pahl. Signal transduction from the endoplasmic reticulum to the cell nucleus. Physiol Rev. 79 (1999) 683-98.

G. Narclai et at. /Reduction of the Endoplasmic Reticulum

289

[38] 1. Braakman, J. Heienius and A. Helenius. Manipulating disulfide bond formation and protein folding in the endoplasmic reticulum. EMBO J, 11 (1992) 1717-22. [39] H.F. Gilbert. The thiol-disulfide exchange. Adv Enzymol Relat Area Mol Biol. 63 (1990) 69-172. [40] M. Bader, W. Muse, D.P. Ballou. C. Gassner and J.C.A. Bardwell. Oxidative protein folding is driven by the electron transport system. Cell. 98 (1999) 217-27. [41] T. Nishikawa, D. Edelstein, X. Liang Du, S. Yamagishi, T. Matsumura, Y. Kaneda, M.A. Yorek, D. Beebe, P.J. Dates, HP. Hamrnes, I. Giardino and M Brownlee. Normalizing mitochondria! superoxide production blocks three pathways of hyperglycemic damage. Nature. 404 (2000) 787-790. [42] B. Lipinski. Pathophysiology of oxidative stress in diabetes mellitus. J Diabetes Complication. 15 (2001) 203-210. [43] Alluru S. Reddi. Riboflavin nutritional status and flavoprotein enzymes in streptozotocin diabetic rats. Biochim Biophys Acta. 882 (1986) 71-76. [44] F.C. Ngkeekwong and Leong L. NG. Two distinct uptake mechanisms for ascorbate and dehydroascorbate in human lymphoblasts and their interaction with glucose. Biochem J. 324 (1997) 22530. [45] G. Nardai, L. Braun, M. Csala, V. Mile, P. Csermely, A. Benedetti, J. Mandl and G. Banhegyi. Protein disulfide isomerase- and protein thiol-dependent dehydroascorbate reduction and ascorbate accumulation in the lumen of the endoplasmic reticulum. J Biol Chem. 276 (2001) 8825-28. [46] A.R. Frand and C. A. Kaiser. Erolp oxidizes protein disulfide isomerase in a pathway for disulfide bond formation in the endoplasmic reticulum. Mol Cell. 4 (1999) 469-77. [47] E.M. Berry, E. Ziv and H. Bar-on. Protein and glycoprotein synthesis and secretion by the diabetic liver. Diabetologia. 19(1980)535-40. [48] J. Young, L.P. Kane, M. Exley and T. Wileman. Regulation of selective protein degradation in the endoplasmic reticulum by the redox potential. J Biol Chem. 268 (1993) 19810-18. [49] R. Mandel, H.J.P. Ryser, F. Ghani, M. Wu and D. Peak. Inhibition of a reductive function of the plasma membrane by bacitracin and antibodies against protein disulfide-isomerase. Proc Natl Acad Sci. 90 (1993)4112-16.

290

Thiol Metabolism and Redox Regulation of Cellular Functions A. Pompella et al. (Eds.) IOS Press, 2002

Analytical Developments in the Assay of Intra-and Extracellular GSH Homeostasis: Specific Protein S-Glutathionylation, Cellular GSH and Mixed Disulphide Compartmentalisation and Interstitial GSH Redox Balance. Ian A. COTGREAVE Division of Biochemical Toxicology, Institute of Environmental Medicine, Karolinska Institute, Box 210, S-l 7177 Stockholm, Sweden. Tel: +4687287654 Fax: +468334467 Ian. cotgreave@Imm. ki. se

1. Backround and Scope Since its discovery in the early part of the 20th century by Sir Fredrick Gowland-Hopkins, glutathione (GSH) has attracted much attention in a variety of disparate scientific fields ranging from basic biochemistry, through toxicology cell biology and physiology and patho-physiology. Despite the myriad of publications and classical treatise [1, 2] detailing the biochemistry of GSH over the years, there are still a number of basic issues begging plausible explanations, particularly with relation to the involvement of the GSH redox buffer in delicate cell regulatory loops in a changing aerobic environment [3,4]. The development of suitable analytical approaches have always been rate-limiting in our understanding of biochemical processes, and the analysis of redox-active species such as GSH presents special requirements on these developments, due to he constant risks of introduction of oxidation artefacts during manipulation of biological material. It is the scope of this paper to detail some of the efforts made in our laboratory to address difficult analytical problems related to the study of intra- and extracellular GSH homeostasis and metabolism. We detail: 1) Attempts to develop an affinity tagging/purification method for high throughput isolation an identification of proteins substrates for reversible S-glutathionylation, an issue of protein redox regulation currently undergoing a renaissance. 2) Immuno-cytochemical methods for the Determination of the intracellular Compartmentalisation of GSH and its oxidised forms (protein-GSH mixed disulphides), and variations in cellular GSH in intact cells with cell cycle and cellular phenotype. 3) Determination of GSH redox states in the extracellular space of intact human tissues using micro-dialysis.

2. Identification of Novel Protein Substrates for S-Glutathionylation during Oxidative Stress and Constitutive Metabolism Our understanding of the complexity of regulation of cellular protein function has begin to focus on redox modifications of protein cysteines by reversible interactions with the soluble GSH redox buffer, so-called reversible protein S-glutathionylation. Indeed, much of the focus of the present symposium resides on the role of such reaction in the overall regulation of cellular function [3,4]. However, emerging from the theoretical into to the practical has

LA. Cotgreave /Assay of Intra- and Extracellular GSH Homeostasis

291

required many decades of often tedious but classical biochemical experimentation. One of the major stumbling blocks to our understanding has been the lack of ability to follow these reactions at the level of individual proteins. This has proven possible with a variety of rather high-copy number soluble proteins [5-7], but has generally been restricted to relationships to intracellular oxidative stress. Similarly, in today's age of high-through put gene analysis, it is the ability to follow organised patterns ofredox modification of proteins that will provide the key to the biochemical "decision making" made in response to the dynamic intracellular redox potential. About 3 years ago our group set about trying to develop an analytical technique which would assist in placing reversible S-glutathionylation reactions on the mantle-piece of cellular regulation. Today this technique is promising to rapidly accelerate both the identification of proteins undergoing reversible S-glutathionylation during oxidative stress, but also allows us the sensitivity to analyse which proteins are undergoing constitutive modification, and perhaps regulation, by this mechanism. The technique may also furnish rapid identification of the position of modification in individual proteins, providing enormous savings as an alternative to more classical sequencing/site-directed mutagenesisbased techniques. The technique centres around the use of an engineered glutaredoxin (Grx) as a "sensor" of S-glutathionylated proteins, selectively reducing these mixed disulphides, and yielding a "fingerprint" of protein thiols once bearing GSH. The bacterial Grx-3 enzyme possesses two mutated cysteines, one in the active site (CHS), ensuring that only GSHprotein mixed disulphides are reduced [8], and an external regulatory thiol (C65Y), which ensures that the enzyme is unable to promote S-glutathionylation via auto-glutathionylation [9]. Initially, cells were treated with diamide and then with excess n-ethylmaleimide (NEM), to remove excess thiols which would complicate eventual tagging of protein thiols liberated by Grx. Cells were washed in such a manner that a functionally low concentration of NEM remains associated with the cells to ensure that artifactual reduction of GSHprotein mixed disulphides does not occur before treatment of the protein with Grx. In the initial application of the methods, validations have been performed in a post-mitochondrial fraction of the cell, largely containing soluble and microsomal proteins. The enzyme was allowed to react with extracted proteins for only a very short period (2-5 minutes), and is started by the addition of GSH (ImM excess over the residual NEM content), GSSG reductase and NADPH, and the use of radio-labelling of intracellular GSH revealed this treatment to facilitate quantitative removal of GSH from protein extracts. The method then requires re-reaction of the stripped protein with NEM-biotin, providing the vital "handle" for subsequent affinity purification on monomeric streptavidin. Avidin eluents were then subjected to 2-dimensional PAGE and randomly selected spots identified by MALDI-TOF analysis (Figure 1). The proteins identified in this manner not only included our old friend actin, as expected with diamide treatment, but a number of other cytoskeltal proteins including factin capping protein and laminin (P40). Importantly, co-ordinated s-glutathionylation of a variety of protein chaperones and stress proteins was also evident, including a variety of protein disulphide isornerase (PDI) precursors, GRP78, HSP90P and HSC 71, as well as antioxidant protein 2. Importantly, a number of cell cycle regulatory proteins underwent forced S-glutathionylation during diamide treatment, including hepatoma-derived growth factor (HDGF-1), nucleoside diphosphate kinase A, CRK-like protein and 14-3-3 protein zeta. A number of novel substrates were also identified amongst enzymes of intermediary metabolism, including lactate dehydrogenase, alpha-enolase and 3-hydroxyacyl-CoA dehydrogenase II.

292

I.A. Cotgreave /Assay oflntra- and Extracellular GSH Homeostasis

Cytosolic proteins stripped of GSHbyC14S,C65Y glutaredoxin- treatment

Substrate identification by tryptic fragmentation and MALDI-TOF

<*

Concentration and 2D gel separation

+

Affinty purification on avidin-sepharose, with biotin elution

Rg. 1. A schematic representation of the glutaredoxln-based affinity chromatographyproteomic analysis of S-glutathlonylated proteins In cells during constitutive metabolism and diamlde-lnduced oxldatlve stress.

Perhaps the biggest step taken with the technique to-date was, however, its application to cells undergoing constitutive aerobic metabolism. This required, however, up-scaling of the start material by an order of magnitude, in order to ensure sufficient material for proteomic analysis from affinity purification. Analysis of the pattern of modification by 2-D gel revealed a number of similarities with the gels obtained from diamide-treated cells, but also revealed important differences, indicating that some proteins which undergo constitutive S-glutathionylation do not neccessarily undergo enhanced modification during oxidative stress, with some perhaps actually undergoing deglutathionylation. Conversely, some species were only modified in diamide-treated cells and clearly not constitutively modified. Thus, a variety of forms of actin, as well as chaperones such as PDI precursors, GRP58, GRP98, HSP60 and 40S ribosomal protein and antioxidant proteins, such as peroxiredoxins 1 and 4, were all identified. Several interesting intermediary proteins were also identified as constitutively S-glutathionylated, including fructose bisphosphate aldolase, inorganic pyrophosphatase, aldose reductase, pyruvate kinase M2 and nicotinamide N-methyl transferase. Another member of the important cell cycle regulatory 14-3-3 protein, iso-form Sigma, was also identified as constitutively Sglutathionylated. Taken together, these comparisons reveal that the regulation of individual protein S-glutathionylation events is not simply dictated by the resident redox status of the GSH pool, but likely to be dependent on a complex interplay between this, the accessibility and reactivity of individual protein thiols and the activity of enzymatic principles directed perusing mixed disulphide formation. Nature is indeed wonderful! These procedures are presently under review for publication [10].

LA. Cotgreave /Assay oflntra- and Extracellular GSH Homeostasis

293

2.1. Future Analytical Developments and Applications In progress of providing these advances in our understanding facilitated by the identification of patterns of S-glutathionylation revealed in resting cells and cells undergoing oxidative stress, the method is now being developed to unleash the full potential of having a "handle" on protein substrates. We are at present exploring a modification whereby the NEM-biotinylated proteins are subjected to tryptic fragmentation from their native states, and the released fragments then subjected to affinity purification and direct separation and sequencing by tandem quadrupole time-of-fiight mass spectrometry (Q-TOF). Thus, combining the speed and sensitivity of the quadrupole mass spectrometry, with the accuracy of the collision sequencing mode of the second mass spectrometry step, it is envisage that far more protein substrates undergoing Sglutathionylation will be quickly identified. It is also envisaged that sequence analysis of the peptides will rapidly reveal which protein cysteines were originally post-translationally modified with cysteine. Opening the doors to technically advanced analytical approaches such as these should then really provide the sensitivity, selectivity, and magnanimity required to begin to chart the subtle nature of redox regulation of cellular function. Who knows, we may even have a separate chapter in future Biochemistry text books dealing with the important issue raised by this meeting!

3. Glutathione and GSH-Protein Compartmentalisations and Population Distributions in Intact Cells One pre-requisite to more fully understanding the subtlety of the regulatory role of the GSH redox buffer in controlling intracellular protein function, is improvement of our knowledge of the disposition of the low molecular weight thiol in cells. During the past decades many analytical approaches have been developed to assay cellular GSH and its oxidised forms, most dependent on some kind of chemical derivatisation of the free thiol, some kind of purification of adducts, followed by quantitation dependent on the physical nature of the adduct. This can be typified by the monobromobimane method typically used by the author and others [11]. These methods do not, however, reveal anything of the intracellular distribution of GSH, nor do they provide any concept of the population distribution of GSH in "homogeneous" cell populations or in mixtures of cells. In the later case, chemical derivatisation, coupled to fluorescence cell sorting (FACS) has provided some insight [12], but these methods are generally hampered by the induction of artefacts due to the dynamic (GSH transferase-dependent) nature of the cell labelling procedures [13]. We recently embarked on the development of a simple assay procedure based on the combination immuno-cytochemical staining of cells with an existing GSH-specific poly-clonal antibody, coupled to fluorescence labelling and analysis either by laser confocal microscopy or FACS. Topgraphical fluorescence imaging of a variety of different cell types labelled with the anti-GSH antibody and an ALEXA-red secondary antibody revealed discontinuous staining throughout the cells, with an apparently low level of GSH in the nuclear area of the cell. In some cell types, a grainy peri-nuclear staining pattern was revealed, which was demonstrated to co-localise with staining for cytochrome oxidase, an established mitochondrial marker. However, it only when the cells were analysed by confocal microscopy, that selective imaging of z-planes cut down through the cell revealed true nature of the discontinuity of staining (Figure 2). The evidence suggest that mitochondria, clustered around the nucleus in these A549 cells, contain very high levels of GSH staining, in comparison with the surrounding cytosol and the nucleus. This mitochondrial pool was also shown to be very stable to depletion, following inhibition of

294

LA, Cotgreuve /Assay oflntra- and Extracellular GSH Homeostasis

cytosolic GSH synthesis. The data also suggest that a cytosolic gradient of GSH might exist, increasing steadily from the apex of the cell down towards the basal contact zone of the cell with the underlying growth surface. The data also demonstrate that in the upper section of the cell, the nuclear and cytosolic staining are near equilibrium. This fascinating insight into the three dimensional distribution of intracellular GSH would therefore suggest that previous estimates of mitochondria! GSH levels, based on isolation of the organelles from intact cells, may grossly underestimate that actual levels of GSH present. This has clear ramifications for researchers working on redox-regulation of mitochondria! function, such as the regulation of permeability transition and release of cytochrome C, so central to the triggering process of many apoptotic events. Top

z-cut 1

z-cut 7

z-cut 3

z-cut 9

z-cut 5

z-cut 11

Bottom

Fig. 2. Laser confocal microscopic analysis of A549 type II bronchial epithelial cell glutathtone. Cells were fixed and stained with a primary polydonal anti-GSH antibody and a secondary ALEXA-red antibody and analysed by confocal microscopy. The images represent 6 of 16 cuts in the z-plane at 0.7 \t spacing, from the upper level of the cells (top left-hand) to the lower levels (bottom right-hand).

In relation to the primary thrust of the present meeting, we also detail application of the antibody staining technique to both tile determination of GSH-protein mixed disulphide distribution in cells, and to the assay of individual cell and organellar GSH levels in a population, as well the determination of population dynamics in GSH-protein mixed disulphide formation. In our preliminary efforts we clearly show the applicability of the antibody staining technique to illuminate the intracellular distribution of GSH bound to protein as a result of diamide metabolism in cells. The data reveal heavy labelling in membrane blebs, punctate staining around the nucleus and, interestingly, heavy nuclear staining. When FACS analysis was combined with labelling, we also demonstrate clear

I.A. Cotgreave /Assay of Intra- and Extracellular GSH Homeostasis

295

population distributions both in GSH levels in a "homogenous" cell population, due perhaps to variation in cellular GSH with position in the cell cycle, as well as in GSHprotein mixed disulphide levels in response to diamide. These initial observations clearly illustrate that we now possess an analytical tool with the resolution to report changes in intracellular and conpartmentalisation and intercellular distribution of GSH, as well as its protein-regulatory "metabolite" at the level of individual cells. 3.1. Future Perspectives and Applications We are currently applying the analytical approaches to several areas of cell biology and toxicology. Firstly, we have been traditionally interested in the role of altered mitochondrial redox balance in early events in apoptosis. Here, it might prove possible to monitor mitochondrial GSH pools in intact cells selectively, perhaps allowing more realistic estimates of the role of altered thiol redox in controlling mitochondrial permeability transitions. The resolution of the cytochemical assay should also greatly facilitate the study of GSH biochemistry in "precious cells", which would otherwise be difficult to assay with conventional chemical techniques. Thus, we are presently focussing on GSH-synthesis-deficient cells, where we are interested to determine the cytosolicmitochondrial GSH distribution patterns, and neuronal stem cells, where fundamental issues of the synthesis of GSH are the issue. Finally, in terms of cellular GSH, were are interested to apply the confocal and FACS analyses to study regional, particularly nuclear, and cellular fluctuations in GSH during the cell cycle. Here the FACS offers opportunities to easily gate cell populations based on specific markers of cell cycle or the DNA content of cells. In terms of charting GSH-protein mixed disulphides visually, we will continue our confocal imaging efforts during oxidative stress, hopefully eventually being able to image constitutive GSH-protein mixed disulphide distribution using immuno-chemical amplification methods. We will also persue the assay of the distribution of GSH-protein mixed disulphides in individual cells and within cell populations, particularly focussing on the current debate within HIV research of aberrant thiol redox balance in infected Tlymphocyte sub-sets. In passing on from this section one is prompted by the population saying that "seeing is believing"!

4. Determination of GSH Redox States in the Extracellular Space of Intact Human Tissues In the previous sections we have dealt with the "life and times" of intracellular GSH and its potential role in regulation of cellular events. However, the GSH redox buffer is also present in the extracellular space, where it may potentially regulate delicate protein redox events on the external surfaces of cells. Present concept, based largely on assays in the circulatory compartment and in tissues obtained from humans and various animal species, entertain the idea of certain tissues excreting GSH into the circulation, i.e. the liver, and certain tissues in the kidney and lung removing it [1]. Despite this, our knowledge of the actual status of the extracellular GSH pool, particularly within the interstitial space of solid human tissues, is extremely limited. This is the consequence both of ethical issues with human subjects, and analytical issues of obtaining relevant interstitial samples and avoiding artefactual redox reactions during sampling. We have recently been studying GSH homeostasis in human skeletal muscle, particularly during localised and systemic trauma, where the bulky muscles rapidly loose

296

l.A. Cotgreave /Assay oflntra- and Extracellular GSH Homeostasis

considerable amounts of their GSH [14]. These findings, as well as data obtained from isolated human skeletal muscle myoblasts and myotubes [15] suggest that muscle may be an important extra-hepatic source of systemic GSH, mobilised by elevations in circulatory stress hormones. In order to continue these studies it was thus necessary to be able to access the muscle interstitial space and obtain estimates of the GSH redox buffer in this. Thus, we utilised a standard micro-dialysis technique, but adapted it to take into account the redox nature of the target molecule. Thus, we developed a twin-rate perfusion protocol, with a fast perfusion rate (2.66 ul/min), lacking to allow full equilibration of GSH and cysteine into the perfusate, but providing a n accurate "redox quota" of reduced to oxidised thiol, and a slow perfusion rate (0.16 ul/min) allowing near complete equilibration of the total thiol content of the interstitium with the dialysate. The quota is then used to extrapolate from this low rate data to determine the absolute levels of reduced GSH in the interstitium.

Q skeletal muscle adipose tissue

O-3

I!160

P-4

E-5

skeletal muscle adipose tissue

5120 0 .$

j j f 80-

c •— ~ §40CO ^

n n

n J-1

B-2

O-3

P-4

E-5

Rg. 3. Free and total glutathione levels In mlcro-dialysate obtained from human skeletal muscle and adipose using a twin perfuslon-rate protocol. Human volunteers were subjected to micro-dialysis at a fast perfusion rate of 2.66 f/l/min, and the free and total GSH levels in the solution determined as detailed in reference 11. This yielded a redox quota which could then be applied to data obtained at a much slower perfusion rate of 0.16 /4/min, allowing optimal equilibration of the thiol into the catheter, where the total glutathione content was determined. The redox quota was then used to calculate the expected levels of reduced GSH in this sample.

Figure 3 details a comparison of the levels of free and total GSH in the interstitial space of skeletal muscle (vastus lateralis) and adipose (periumbilical, sub-cutaneous) of 5 healthy volunteers. Three observations are striking. First, skeletal muscle interstitial levels are always strikingly greater than those present in adipose interstitium. The levels of free GSH were also considerably above the respective levels of free GSH in the central plasma

I.A. Cotgreave /Assay of Intra- and Extracellular GSH Homeostasis

297

compartment. This striking difference was not evident in the cysteine redox pool, which was rather equivalent in both tissue interstitiae. Secondly, there is a large variation between individuals in the skeletal muscle interstitial GSH levels. This variation was not as great in the cysteine pools and there appeared no correlations between the variations in the individual levels or redox states of these two thiols, in eithe of the tissues. Lastly, GSH redox lies considerably on the reduced side in skeletal interstitium, more so than in the adipose tissue interstitium or in the central plasma water compartment. This was the opposite to the case with the interstitial cycteine redox state assayed in the interstitiae. Taken together, these data strongly support the concept of a unique pool of reduced GSH present in the interstitial space of human skeletal muscle, maintained by active excretion from the myotubes. Given the bulk of this tissue (up to 40% of our body weight in healthy individuals), this pool of GSH may reflect a fundamental function of skeletal muscle as an extra-hepatic source of GSH in the body. The data also strongly suggest that, when one is rationalising attempts to regulate cellular function by application thiols and/or disulphides, simple extrapolation of arguments to the "normal" plasma levels of reduced and oxidised GSH and cysteine may not apply. Clearly there are cells in the body exposed to considerably higher extracellular GSH levels than previously considered. Knowledge of the correct levels and redox status of GSH in the extracellular space will greatly assist attempts to define the potential role of this redox buffer in regulating membrane protein function at the external cell surface by reversible protein Sgluathionylation reactions 4.1. Future Developments and Applications In terms of GSH in the muscle interstitial space, we are particularly interested to apply the two-rate perfusion technique to probe for fluctuations in response to localised trauma, injury and infammation, as well as to systemic trauma as a result of post-operative shock and intensive care. Although difficult to perform clinically, these manipulations may provide a better insight into the systemic regulation of GSH availability and utilisation, both in health and disease. We are also interested to adapt the micro-dialysis technique to the perfusion of several other tissues, where oxidative stress is proported to be central to the pathophysiological causes/consequences of disease. The motto here is definitely "glutathione has a life on the outside of the wall as well"! Another focus of the research is to extend our analyses to other redox-active molecules in skeletal muscle and adipose interstitiae, such as ascorbate and vitamin E. This will provide us with a more accurate picture of the total antioxidative capacity of tissue extracellular spaces.

A cknowledgements The author acknowledges the contributions of his co-workers in the various projects: Christina Lind (Karolinska Institute, KI), Robert Gerdes (Retired), Arne Holmgren (KI), Ina Schuppe-Koistinen (AstraZeneca, AZ) and Helena Brockenhuus von Lowenhielm (AZ), (Sglutathionylation assays). Therese Soderdahl (KI), George Bocsfoldi (AZ), Marie Enochsson (KI) and Mathias Lundberg (KI), (GSH cellular topology). Michail Tonkonogi (KI) and Jan Henriksson (KI), (Interstitial GSH redox states). Financial support is acknowledged from The Swedish Medical Research Council (VR), The Swedish Centre for Sports Research (CIF), the Karolinska Institute and AstraZeneca.

298

I.A. Cotgreave /Assay oflntra- and Extracellular GSH Homeostasis

References 1. Meister A and Anderson M (1983). Glutathione. Ann. Revs Biochemistry, 52, 711 -760. 2. Orrenius S, Ormstad K, Thor H and Jewel SA (1983). Turnover and function of glutathione studied with isolated hepatic and renal cells. Fed. Proc., 42,3177-3188. 3. Cotgreave IA and Gerdes GR (1998). Recent trends in glutathione biochemitry: Glutathione -protein interactions: A molecular link between oxidative stress and cell proliferation? Biochem. Biophys. Res. Comms., 242,1-& 4. Klatt P. Lamas S (2000). Regulation of protein function by S-glutathiolatk>n in response to oxidative and nitrosative stress. European Journal of Biochemistry, 267,4928-44. 5. Chai YC et al (1991). Identification of an abundant S-thiolated rat liver protein as carbonic anhydrase III; characterization of S-thiolation and dethiolatton reactions. Archives of Biochemistry & Biophysics. 284, 270278. 6. Schuppe-Koistinen I, MokJeus P, Bergman T and Cotgreave IA (1995). Reversible S-gluathionyiation of human endothelial cell actin accompanies a structural rearragement of the cytoskeleton. Endothelium, 3, 301-308. 7. Schuppe-Koistinen I, Moldeus P, Bergman T and Cotgreave IA (1994). S-thiolation of human endothelial cell glyceraldehyde-3-phospnate dehydrogenase after hydrogen peroxide treatment. European Journal of Biochemistry. 221,1033-1037. 8. Bushweller JH. Aslund F. Wuthrich K. Holmgren A (1992). Structural and functional characterization of the mutant Escherichia coli glutaredoxin (C14—S) and its mixed disulfide with glutathione. Biochemistry. 31, 9288-9293. 9. LJnd C. Gerdes R. Schuppe-Koistinen I. Cotgreave IA (1998). Studies on the mechanism of oxidative modification

of

human

glyceratdehyde-3-phosphate

dehydrogenase

by

glutathione:

catalysis

by

glutaredoxin. Biochemical & Biophysical Research Communications. 247, 481-486. 10. Lind C, Gerdes R, Hamnell, Y, Schuppe-Koistinen I, Brockenhuus von Lowenhielm H, Holmgren A and Cotgreave IA. Identification of S-glutathionylated cellular proteins during oxidative stress and constitutive metabolism by affinity purification and proteomic analysis. Nature Biotechnology under review. 11. Cotgreave IA and Moldeus P (1986). Methodologies for the application of of biological systems. Journal of Biochemical & Biophysical Methods. 13, 231-249. 12.

Peterson JD. Herzenberg LA. Vasquez K. Waltenbaugh C (1998). Glutathione levels in antigen-presenting cells modulate Th1 versus Tn2 response patterns. Proceedings of the National Academy of Sciences of the United States of America. 95(6), 3071-3076.

13. Briviba K, Fraser G, Sies H and Ketterer B (1993). Distribution of monchlorobimane-gluathione conjugate between nucleus and cytosol in isolated hepatocytes. Biochem J., 294, 631-633. 14. Hammarqvist F, Luo JL, Cotgreave IA, Andersson K and Wemerman J (1997). Skeletal muscle glutathione is depleted in critically ill patients. Critical Care Medicine. 25, 78-84. 15. Cotgreave IA, Goldschmkft L, Tonkonogi M and Svensson M (2002). FASEB J Express, 16, 435-437.

Thiol Metabolism and Redox Regulation of Cellular Functions A. Pompella et al. (Eds.) IOS Press, 2002

Signalling Potential and Protein Modifying Ability of Physiological Thiols U. MURA, M. CAPPIELLO, P.O. VILARDO, I. CECCONI, M. DAL MONTE and A. DEL CORSO Dip. di Fisiologia e Biochimica, Universita di Pisa, via S. Maria, 55 - 56100 Pisa, Italy 1. Introduction Oxidative stress is one of the most deleterious conditions for cell systems. A number of oxidative modification processes involving highly relevant biological molecules can be seen as primary events in the development of pathological states which may lead, depending on the severity of the oxidative insult, to cell death [1-5]. Among others, thiols represent an important class of molecules which are easily involved in redox processes. This because of their generally high susceptibility to oxidation which may occur in biological systems at relatively high rate without the assistance of enzymatic systems. Moreover, the potential reversibility of at least the initial steps of thiol oxidation (i.e. thiol to disulfide) and the presence of thiols on proteins gives to these functional groups the feature to directly connect oxidative phenomena with the molecular machinery responsible for the metabolic control. Another emerging aspect of thiol/disulfide interconversion is the paradoxical prooxidant ability of thiol molecules [6-11] whose oxidation, linked to CVFbOa interconversion, may be associated to cell signalling phenomena [12-16]. Indeed, the scavenging action exerted by low molecular weight thiols in counteracting oxidative stress conditions can be amplified via the S-thiolation of enzymes [17-19]. However, whether S-thiolation is a way to control enzyme activity or is the unavoidable consequence of the reactivity of thiols is still a matter of debate, essentially because of the lack of specialised enzymes able to catalyze thiolation/de-thiolation reactions. In addition, even though the reaction rates of non-enzymatic S-thiolation processes are sufficiently high to be compatible with cellular functions, the problem linked to the required specificity of the modification stands. In this regard, however, it should be considered that the target molecules are highly structurally organized. Thus the restraint in the accessibility of specific thiolating agents into the thiolation site, combined with their peculiar effects on the target protein, may impose sufficient restrictions to confer to the process features of specificity. However, it is also true that the body of data on enzymes able to intervene into the protein thiol/disulfide interconversion is growing [20-23], so that the lack of enzymatic driven S-thiolation processes may only be a temporary lack of knowledge. Making use of aldose reductase (alditol: NADP+ oxidoreductase, EC 1.1.1.21) (ALR2) as a protein model, our studies on thiol mediated oxidation processes confirm the ability of different thiols to determine, via S-thiolation, specific modification of the enzyme. Moreover, they support the view of physiological thiols as components of a machinery able to promote and propagate redox signals. In this regard, both cysteine and Cys-Gly are able to promote in a

300

U. Mura et al. /Signalling Potential and Protein Modifying Ability

catalytic fashion the mobilization of GSH into the thiol/disulfide interconversion machinery. 2. Physiological Thiols as Promoters and Propagators of Reactive Oxygen Species Even though all thiol compounds have the potential to reduce molecular oxygen, their effectiveness in such a performance is quite dependent on their structural features [10, 24]. Unlike what observed for GSH which is rather stable to oxidation, Cys and Cys-Gly in the presence of transition metal ions are able to promote oxygen activation. Thus, while disulfides are generated, reactive oxygen species (ROS) such as superoxide ion, hydroxyl radical and hydrogen peroxide can be easily produced in the presence of this two physiological thiols [611]. The ability to generate ROS may confer to these molecules the duty for a more general cell signalling action. Such an effect, may be apparently limited by the usually rather low level of Cys and Cys-Gly in cell systems. However it can be significantly amplified in the presence of GSH which is driven into the redox process by the relative high susceptibility to oxidation of both Cys and Cys-Gly.

0.8

0.4

Fig. 1. [Iron-EDTA/Cysl-induced oxidation of GSH. 1 mM GSH was incubated at 37°C in 100 mM sodium phosphate buffer pH 6.8 in the presence of 0.3 mM FeSOt, 0.9 mM EDTA and 0 (filled circles), 0.1 (triangles), 0.2 (squares), or 0.5 (diamonds) mM cysteine. Control (open circles) contained neither Cys or FeSO4.

In fact, a massive mobilization of GSH is observed in the presence of Ox and transition metal ions, which are poorly efficient in promoting GSH oxidation, when either Cys or CysGly are present The effect exerted by different level of Cys on GSH oxidation is shown in

V. Mum et al. /Signalling Potential and Protein Modifying Ability

3()1

Fig.l. While the rate of GSH oxidation is significantly affected by the level of Cys or CysGly, the extent of oxidation, which is essentially complete, appears unaffected by the level of the two thiols. Even though the generation of ROS, following the prooxidant action of Cys and Cys-Gly, may contribute to GSH oxidation, the analysis of reagents, intermediates and end products on Cys/GSH mixtures undergoing oxidation revealed the presence of the mixed disulfide cysteinyl-glutathione as a relevant intermediate in the oxidation process. This result indicates that the increase in the rate of GSH oxidation induced by Cys occurs through transthiolation processes. Thus, in oxidative conditions, the generation of mixed-disulfides leads to a recycle of Cys, and likely Cys-Gly, which comes to be catalysts of the process. So, even though usually present in the cell systems at rather low concentrations, Cys and Cys-Gly can play a role in the non-enzymatic involvement of GSH in the antioxidant defences. Because of their effectiveness in enhancing the overall thiol oxidation rate in the presence of metal catalyzed oxidation systems, Cys and Cys-Gly may be a relevant factor in the antioxidant action of the cell. Indeed, their low reduction potentials as compared to GSH, may amplify the sensitivity of the antioxidant defence system, by enhancing its responsiveness to counteract the oxidative insult. While the increased responsiveness to ROS promoted by Cys and Cys-Gly, can favourable play against oxidative conditions, the prooxidant action of these thiol compounds is difficult to be seen as a part of an antioxidant mechanism. Nevertheless the mobilization of GSH by Cys and Cys-Gly in a non-enzymatic red/ox process, may still be seen as a control system of oxidative phenomena able to modulate the cellular oxygen tension. Even though no specific study has been performed to verify such a potential role of GSH, the remarkable increases in the GSSG/GSH ratio, observed in different ocular lens models subjected to hyperbaric oxygen treatment [25-27] makes this hypothesis worth of further investigation. It is obvious that the higher sensitivity of cell thiols to oxidants requires an increase in the consumption of cell reducing equivalents in order to keep the proper thiol/disulphide ratio. Such a need, which essentially means availability of NADPH, can be fulfilled even because of the apparent concerted action of S-thiolation on metabolically linked enzymes [18], whose function may be directed in counteracting oxidative stress. This would be realized both by channelling metabolic resources toward the synthesis of NADPH and by reducing NADPH consumption [19], as in the case of aldose reductase, the protein target of our studies on Sthiolation.

3. Physiological Thiols as Protein Modifying Agents Bovine lens aldose reductase, appeared to be a very useful system to show how efficient and at the same time specific the action of different physiological thiols is. The susceptibility of ALR2 to S-thiolation processes induced by physiological and non physiological disulfides (or by the corresponding thiols in oxidative conditions) is to be ascribed to the ability of Cys298 to form mixed disulfides with low molecular weight thiols. Indeed, among the seven Cys residues present on the enzyme, Cys298, because of its accessibility, is the only residue susceptible of S-thiolation; this irrespectively of the nature of the thiol used as thiolating agent. However, ALR2 forms carrying different thiols linked to Cys298 differ both in terms of kinetic and structural properties. A comparison of the specific activities and susceptibility to inhibition by Sorbinil of different enzyme forms modified at level of Cys298 by different physiological disulfides and non-physiological thiol agents is shown in Fig.2.

302

U. Mum el al. /Signalling Potential and Protein Modifying Ability

Fig. 2. Sensitivity to Sorbinil Inhibition of different thlot-modlftod ALR forms. Different ALR2 forms were assayed using D,L-glyceraldehyde as substrate in the absence (closed bars) and in the presence (open bars) of 10 (M Sorbinil. GS-ALR2: glutathione-modified ALR2; 2ME-ALR2: 2mercaptoetanol-rnodified ALR2; TG-ALR2: monothioglycerol-modifted ALR2; Cys-ALR2: cysteinemodified ALR2; Cys-Gly-ALR2: cysteinylglycine-modifie ALR2; Cbm-ALR2: carboxymethylated ALR2; Cbam-ALR2: carboxyamktomethylated ALR2. Different enzyme forms were prepared as previously described [28-31]

The only feature common to all the thiolated ALR2 forms as well as to the human recombinant C298S mutant [32], is a reduced or complete loss of susceptibility to the inhibitory action of aldose reductase inhibitors. Moreover, it appears that, with the only exception of the glutathionyl modified enzyme (GS-ALR2), the engagement of the thiol group of Cys298 in a S-S or S-C bond, or the exchange of the SH with an OH group, as is the case of the C298S ALR2 mutant, does not compromise and actually in most cases increases the glyceraldehyde reducing activity of the enzyme. It is difficult so far to envisage the rationale for the apparent activation observed in the majority of the different modified enzyme forms. On the contrary, the reduced specific activity displayed by GS-ALR2 is likely due to the protrusion of the Y-glutamyl moiety of the linked glutathionyl residue into the active site. Such an allocation of the glutathionyl moiety determines an interaction with Tyr48, which results impaired in its function of proton donor in the redox mechanism of the enzyme [33-35]. Indeed, energy minimization and molecular dynamic approaches reveal a degree of interaction between GSH and the enzyme sufficient to define for ALR2 a GSH-binding site [36], This would find support into the remarkable effectiveness of ALR2 in the reduction of the adduct GSHrhydroxynonenal [37]. Thus, despite targeting to the same reactive group, GSH appears unique as S-thiolating agent in reducing the ALR2 activity. Such a specificity is accompanied

U. Mum et al. /Signalling Potential and Protein Modifying Ability

303

by a remarkable stability of GS-ALR2 with respect to other S-thiolated enzyme forms. In fact, unlike what observed for GS-ALR2, both Cysteine-modified ALR2 and Cys-Gly-modified ALR2 [30] and likely 2ME-ALR2 [29] easily tends to an intramolecular rearrangement which leads to the formation of a disulfide bond between Cys298 and Cys303. Such a structural rearrangement which determines a complete inactivation of the enzyme would appear to minimize in terms of function, even though indirectly, the above mentioned specificity associated to the GS-ALR2 formation. Nevertheless, it is strongly indicative, in general terms, of the special structural features acquired by the target protein depending on the nature of the modifying thiol. It is worth to note, in this regard, that only the glutathionyl-modified ALR2 form was found so far in intact cultured lenses subjected to oxidative stress by hyperbaric oxygen treatment [25]. This enzyme form, which appears rather sensitive to the catalytic action of glutaredoxin (Fig. 3), can be easily reduced back to the native enzyme as it occurs in intact lens cultured in normobaric air conditions, after ceasing the oxidative stress [38].

Fig. 3. Time course of GS-ALR2 reactivation. GS-ALR2 (3jiM) was incubated in 10 mM sodium phosphate buffer pH7.0 at 37°C in the presence of 0.5 mM GSH alone (squares) or in the presence of an equimolar amount of rat liver glutaredoxin (triangles). Control (circles) contained neither GSH nor glutaredoxin.

The question whether ALR2 function includes the reversible S-glutathionylation of the enzyme as a regulatory event cannot be answered at this point. However, it is suggestive the fact that another enzyme, whose NADPH-consuming activity is impaired by protein Sthiolation, can be included in the above mentioned enzymatic machinery [18] apparently devoted to save cell reducing power for antioxidant purposes.

References 1) J.F.Jr. Keaney and J. Vita, Atherosclerosis, oxidative stress, and antioxidant protection in endothelium-derived relaxing factor action. Prog. Cardiovasc. Ois. 38 (1995) 129-154. 2) C.D. Smith, J.M. Carney, P.E. Starke-Reed, C.N. Oliver, E.R. Stadtman, R.A. Floyd and W.R. Markesbery, Excess brain protein oxidation and enzyme dysfunction in normal aging and in Alzheimer disease, Proc. Natl.

304

U. Mum et al. / Signalling Potential and Protein Modifying Ability

Acad. Set. USA 88 (1991) 10540-10543. 3) R. Adelman, R.L Saul and B.N. Ames, Oxidative damage to DMA: relation to species metabolic rate and life span, Proc. Natl. Acad. Sci USA 85 (1988) 2706-2710. 4) E.R. Stadtman and C.N. -Oliver, Metal-catalyzed oxidation of proteins. Physiological consequences, J. Biol. Chem. 266 (1991) 2005-2008. 5) B. Halliwell, Oxidants and human disease: some new concepts, FASEB J. 1 (1987) 358-364. 6) D.A. Rowley and B. Halliwell, Superoxide-dependent formation of hydroxyl radicals in the presence of thiol compounds, FEBS Lett. 138 (1981) 33-36. 7) A.A. Stark, E. Zeiger and D.A. Pagano, Glutathione metabolism by Y-glutamyttranspeptidase leads to HpkJ peroxidation: characterization of the system and relevance to hepatocarcinogenesis, Carcinogenests 14 (1993) 183-189. 8) P. Perego, A. Paolicchi, R. Tongiani, A. Pompella, P. Tonarelli, N. Carenini, S. RomaneHi and F. Zunino, The cell-specific anti-proiferative effect of reduced glutathione is mediated by y-glutamyl transpeptidase-dependent extracellular pro-oxidant reactions, Int. J. Cancer 71 (1997) 246-250. 9) R. Drozdz, C. Parmentier, H. Hachad, P. Leroy, G. Siest and M. Wellman, Gamma-gKitamvttransferasedependent generation of reactive oxygen species from a glutathioneAransferrin system, Free Rad. Biol. Med. 25(1998)786-792. 10) M. Enoiu, H. Haberkane, J.F. Salazar, P. Leroy, J. Groffen, G. Siest and M. Wellman, Evidences for the prooxidant effect of y-glutamyftranspeptidase-related enzyme, Free Rad. Biol. Med. 29 (2000) 825-833. 11) A. Del Corso, P.G. vllardo, M. Cappielk), I. Cecconi, M. Dal Monte, D. Barsacchi and U. Mura, Physiological thiols as promoters of glutathione oxidation and modifying agents in protein S-thblation, Arch. Biochem. Biophys. 397 (2001) 392-398. 12) M. Ohba, M. Shibanuma, T. Kuroki and K.J. Nose, Production of hydrogen peroxide by transforming growth factor-beta 1 and its involvement in induction of EGR-1 in mouse osteoblastic cells, J. Cell Biol. 126 (1994) 1079-1088. 13) B. Meier, H.H. Radeke, S. Selte, M. Younes, H. Sies, K. Resch and G.G. Habermehl, Human fibroWasts release reactive oxygen species in response to interleukin-1 or tumour necrosis factor-alpha, Biochem. J. 263 (1989)539-545. 14) Y-S. Bae, J-Y. Sung, OS. Kim, Y-J. Kim, K.C. Hur, A. Kazlauskas and S.G. Rhee, Platelet-derived growth factor-induced H2O2 production requires the activation of phosphatidylinositol 3-kinase, J. Biol. Chem. 275 (2000)10527-10531. 15) H.P. Monteiro and A. Stem, Redox modulation of tyrosine phosphorylation-dependent signal transduction pathways. Free Rad. Biol. Med.1 (1996) 323-333. 16) Y.J. Suzuki, H.J. Forman and A. Sevanian, Oxidants as stimulators of signal transduction, Free Rad. Biol. Med. 22(1997)269-285. 17) D.M. Ziegter, Role of reversible oxidation-reduction of enzyme thiol-disulfides in metabolic regulation, Annu. Rev. Biochem. 54 (1985) 305-329. 18) R. Brigelius, Mixed disulfides: Biological functions and increase in oxidative stress. In: H. Sies (ed.), Oxidative stress. Academic Press, New York, 1985, pp.243-272. 19) A. Del Corso, M. Cappiello and U. Mura, Thiol dependent oxidation of enzymes: the last chance against oxidative stress, Int. J. Biochem. 26 (1985) 745-750. 20) K. Axelsson and B. Mannervik, An essential role of cytosolic thioltransferase in protection of pyruvate kinase from rabbit liver against oxidative inactivation, FEBS Lett. 7 (1983) 114-118. 21) T. Terada, H. Maeda, K. Okamoto, T. Nishinaka, T. Mizoguchi and T. Nishinara, Modulation of glutathione Stransferase activity by a thiol/disulfide exchange reaction and involvement of thioltransferase, Arch. Biochem. Biophys. 300 (1993) 495-500. 22) T. Terada, Characterization of multiple acid phosphatases in bovine liver cvtosol and lysosome. Inactivation of cytosolic enzymes by disulfides and its redox regulation by thioltransferase, Int. J. Biochem. CeH Biol. 29 (1997) 985-992.

U. Mum ct al. /Signalling Potential and Protein Modifying Ability

305

23) F. Qiao, K. Xing and M.F. Lou, Modulation of lens glycolytic pathway by thioltransferase, Exp. Eye Res. 70 (2000) 745-753. 24) R.E. Benesch, R. Benesch, The acid strenght of the -SH group in cysteine and related compounds, J. Am. Chem. Soc. 77 (1955) 5877-5881. 25) M. Cappiello, P.G. Vilardo, I. Cecconi, V. Leverenz, F.J. Giblin, A. Del Corso and U. Mura, Occurrence of glutathione-modified aldose reductase in oxidatively stressed bovine lens, Biochem. Biophys. Res. Commun. 207(1995)775-782. 26) F.J. Giblin, L. Schrimscher, B. Chakrapani and V.N. Reddy, Exposure of rabbit lens to hyperbaric oxygen in vitro: regional effects on GSH level, Invest. Ophthalmol. Vis. Sci. 29 (1988) 1312-1319. 27) F.J. Giblin, V.A. Padgaonkar, V.R. Leverenz, L-R. Lin, M.F. Lou, N.J. Unakar, L Dang, J.E.Jr. Dickerson and V.N. Reddy, Nuclear light scattering, disulfide formation and membrane damage in lenses of older giunea pigs treated with hyperbaric oxygen, Exp. Eye Res. 60 (1995) 219-235. 28) M. Cappiello, M. Voltarelli, M. Giannessi, I. Cecconi, G. Camici, G. Manao, A. Del Corso and U. Mura, Glutathione dependent modification of bovine lens aldose reductase, Exp. Eye Res. 58 (1994) 491-501. 29) M. Giannessi, A. Del Corso, M.Cappiello, M. Voltarelli, I. Marini, D. Barsacchi, D. Garland, M. Camici and U. Mura, Thiol-dependent metal catalyzed oxidation of bovine lens aldose reductase. I. Studies on the modification process, Arch. Biochem. Biophys. 300 (1993) 423-429. 30) P.G. Vilardo, A. Scaloni, P. Amodeo, C. Barsotti, I. Cecconi, M. Cappiello, B. Lopez Mendez, R. Rullo, M. Dal Monte, A. Del Corso and U. Mura, Thiol/disulfide interconversion in bovine lens aldose reductase induced by intermediates of glutathione turnover, Biochemistry 40 (2001) 11985-11994. 31) A. Del Corso, M. Dal Monte, P.G. Vilardo, I. Cecconi, R. Moschini, S. Banditelli, M. Cappiello, L. Tsai and U. Mura, Site- specific inactivation of aldose reductase by 4-hydroxynonenal, Arch. Biochem. Biophys. 350 (1998) 245-248. 32) J.M. Petrash, T.M. Harter, C.S. Devine, P.O. Olins, A. Bhatnagar, S-Q. Liu and S.K. Srivastava, Involvement of cysteine residues in catalysis and inhibition of human aldose reductase, J. Biol. Chem. 267 (1992) 2483324840. 33) M. Cappiello, P. Amodeo, B. Lopez Mendez, A. Scaloni, P.G. Vilardo, I. Cecconi, M. Dal Monte, S. Banditelli, F. Talamo, V. Micheli, F.J. Giblin, A. Del Corso and U. Mura, Modulation of aldose reductase activity through Sthiolation by physiological thiols, Chem-Biol. Inter. 130-132 (2001) 597-608. 34) K.M. Bohren, C.E. Grimshaw, C-J. Lai, D.H. Harrison, D. Ringe, G.A. Petsko and K.H. Gabbay, Tyrosine-48 is the proton donor and histidine-110 directs substrate stereochemical selectivity in the reduction reaction of human aldose reductase: enzyme kinetics and crystal structure of the Y48H mutant enzyme, Biochemistry 33 (1994)2021-2032. 35) D.K. Wilson, I. Tarle, J.M. Petrash and F.A. Quiocho, Refined 1.8A structure of human aldose reductase complexed with the potent inhibitor zopolrestat, Proc. Natl. Acad. Sci. USA 90 (1993) 9847-9851. 36) K.V. Ramana, B.L. Dixit, S. Srivastava, G.K. Balendira, S.K. Srivastava and A. Bhatnagar, Selective recognition of glutathiolated aldehydes by aldose reductase, Biochemistry 39 (2000) 12172-12180. 37) S. Srivastava, A. Chandra, A. Bhatnagar, S.K. Srivastava and N.H. Ansari, Lipid peroxidation product, 4hydroxynonenal and its conjugate with GSH are excellent substrates of bovine lens aldose reductase, Biochem. Biophys. Res. Commun. 217 (1995) 741-746. 38) M. Cappiello, P.G. Vilardo, V. Micheli, G. Jacomelli, S. Banditelli, V. Leverenz, F.J. Giblin, A. Del Corso and U. Mura, Thiol disulfide exchange modulates the activity of aldose reductase in intact bovine lens as a response to oxidative stress, Exp. Eye Res. 70 (2000) 795-803.

306

Thiol Metabolism and Redox Regulation of Cellular Functions A. Pompella el al. (Eds.) IOS Press. 2002

Redox Signaling and the Map Kinase Pathways Martine TORRES1 & Henry Jay FORMAN2 'Childrens Hospital Los Angeles Research Institute, Department of Pediatrics, Keck School of Medicine, University of Southern California, Los Angeles, CA 90027, USA, and 2 Department of Environmental Health Sciences, University of Alabama at Birmingham, AL, 35294, USA.

1. Introduction Cells respond to their environment and carry out specific functions through highly organized networks of signaling pathways that are tightly regulated and encompass kinases, phosphatases, adapter and scaffold proteins, phospholipases and others that produce or respond to small diffusible molecules called second messengers. While cyclic AMP and calcium are classical second messengers, evidence has recently accumulated in support of a role for reactive oxygen species (ROS) as second messengers and critical modulators of protein phosphorylation and gene transcription. Although cells contain antioxidant enzymes and several reducing systems such as glutathione and thioredoxin (Trx), ROS can produce transient changes in the cellular redox state and, in particular in the redox state of cysteinyl thiols, which can affect the activity, protein-protein and DNA-protein interactions of enzymes and transcription factors. The discovery that ligand-stimulated production of HaO2 was required for the mitogenic response to PDGF in vascular smooth muscle cells [1] and that EOF induced production of H2O2 [2] led to a flurry of studies showing that ROS production by growth factors, cytokines and ligands for G-protein coupled receptors, possibly through oxidases related to the phagocyte NADPH oxidase, represents a common feature of multiple signaling pathways (Figure 1) and is essential for downstream propagation of signals. Several extensive reviews have recently been published in this area [3-6]. Here, we will concentrate on the role of ROS in the activation of the mitogen-activated proteins (MAP) kinases. 2. The Map Kinases and their Activation Modules The MAP kinases are major components of signaling pathways that control proliferation, differentiation, embryogenesis and cell death and belong to a large family of proline-directed serine/threonine kinases. They are the terminal kinases of three-tiered kinase modules that are activated by a variety of stimuli acting through diverse receptor families and they exhibit a high level of conservation from yeast to man. Full activation of the MAP kinases requires dual phosphorylation of a tyrosine and a threonine within a TxY motif in the activation loop. This step is the most specific of the cascade and is performed by dual specificity kinases, the MAP kinase kinases (MAPKK), which are themselves activated by phosphorylation by MAP kinase kinase kinases (MAPKKK) (Figure 2).

M. Torres and H.J. Forman / Redox Signaling and the Map Kinase Pathways

307

PKC

Fig. 1. Redox Signaling. Transient production of ROS by NADPH oxidases may directly affect the redox state of protein thiols in signaling molecules as demonstrated for PTPs and others while the mechanisms altering the activity of others such as the MAP kinases is either indirect or unknown.

MAPKKK

MAPKK

MARK

TF

Elkl Sap-1

c-Jun EJk-1 ATF2

MEF2C Elk-1 ATF2 CHOP

MEF2C Sap-1

Fig. 2. MAP klnase pathways In mammalian cells.

The MAPKKK are activated either by phosphorylation or by interaction with a small GTPase of the Ras or Rho family. Four subsets of MAP kinases activated by separate kinase cascades have been identified in mammalian cells thus far: the classical extracellular signal regulated

308

M. Torres and H.J. Forman / Redox Signaling and the Map Kinase Pathways

kinases (ERK, TEY) ERK1/2, the c-jun N terminal kinases (INK, TPY) JNK1/2/3, the p38 MAP kinases (TOY) and ERK5 (TEY), also called Big MAPK because its high molecular mass, in contrast to the 38-45 kDa of the other MAP kinases. For more detailed reviews, see [7-10]. Activation of the p44ERKl and p42ERK2 by mitogenic stimulation and receptor tyrosine kinases was first recognized and the ERK module, which is composed of ERK1 or ERK2, MEK1 or MEK2 and Raf isoforms is the best characterized, although some questions remain as to the respective role of the c-Raf and B-Raf isoforms [11] and the complex mechanisms of activation by the isoforms of the small GTPase p21 Ras, which provides the link between the ERK module and the receptor at the plasma membrane [12, 13]. The ERKs are also activated by ligands for G-protein linked receptors and non-receptor tyrosine kinases have been shown to be involved [14]. The INK and p38 MAP kinases are primarily activated by cellular stresses such as inflammatory cytokines but also UV, y-irradiation and others. MKK4/SEK1 and MKK7 are the MAPKK for JNK. Interestingly, MKK4 preferentially phosphorylates Tyr-185 while MKK7 targets only Thr-185. Synergistic activation was demonstrated in vitro and in knockout mouse models, indicating that both MKK4 and MKK7 may be required for full activation of the JNKs by specific stimuli [15]. MKK3 and MKK6 are the dual specificity kinases for p38 MAPK and appear to selectively phosphorylate particular p38 isoforms, with MKK3 acting on the a and {J isoforms. At the levels of the MAPKKKs, many kinases have been identified in overexpression experiments that can activate either or both JNK and p38MAPK (Figure 2). ERKS is activated by MEK5 and the upstream kinase is not clearly identified, although MEKK2 and MEKK3 have been implicated [10, 16]. Further work will be required to clearly ascertain the specific kinases for each cascade. In addition, the discovery of scaffold proteins such as the yeast Ste5p that plays a critical role in the activation of the yeast mating pathway by pheromone and selectively tethers the MAPK module in the pathway raises new question. No mammalian counterpart to SteSp has been identified so far but several proteins have been shown to play a role as MAP kinase scaffold such as MIP-1 [17], the beta-arrestins [18] or the misnamed JNK inhibitory proteins (JIPs) [19]. The existence of such anchoring proteins may represent molecular facilitators of MAP kinase activation and may prevent in vivo the extensive cross-talk between pathways that has been implied by overexpression experiments. The activation of the MAP kinases results in proline-directed phosphorylation of various cytosolic and nuclear substrates and despite phosphorylation of a similar sequence (xS/T-P), specificity is observed. Several cytoskeletal proteins, protein kinases and transcription factors are substrates for the ERK or p38 MAPK while only a few transcription factors are targets for the JNKs, although it has been shown that JNKs phosphorylate the Bcl2 proteins [20]. In fact, all substrates contain specific conserved sequences that are distinct from the phosphoacceptor site and are necessary for recognition/binding and phosphorylation by MAP kinases, the delta domain of c-jun being a prototype. These sequences bind to specific sites on the MAP kinases, providing strong specificity [21]. The MAP kinases play a central role in gene regulation by directly phosphorylating distinct transcription factors or by activating other kinases in the cytoplasm and the nucleus. Best studied is the regulation of AP-1 activity by the MAPKs, which increase the transcription of the c-jun and c-fos genes and phosphorylate the newly synthesized components of the AP-1 complex [22]. Nevertheless, one should remember that induction of a particular gene often requires activation of more than one transcription factor, which may be modified by several MAP kinases and by other signaling pathways. Thus, activation of several MAP kinases by receptor/ligand interactions may converge in the nucleus

M. Torres and H.J. Forman / Redox Signaling and the Map Kinase Pathways

309

at the level of transcription, making it difficult to assess the role of each cascade on gene regulation. For example, regulation of the c-jun gene requires ATF-2, c-Jun and MEF2, which are all substrates for MAPKs while SRF, TCF/Elkl and CREB are involved in c-fos regulation, thereby requiring integration of signals transduced by several MAP kinases acting on more than one regulatory elements of the promoter, in concert with other kinases [23]. The duration and extent of activation of the MAPK is governed by the equilibrium between the activity of kinases and phosphatases. In vitro, the serine/threonine phosphatase PP2a and tyrosine phosphatases can dephosphorylate the MAP kinases and some evidence suggests that they may also play a role in vivo, especially for ERK1/2 under conditions where activation is very transient, such as after stimulation by fMLP in neutrophils [24]. In addition, a large family of dual specificity phosphatases (DSP), also called MAP kinase phosphatases (MKP), has been identified as inducible proteins, exhibiting specificity for the MAP kinases [25]. DSP contain a signature motif that is present in all protein tyrosine phosphatases, a property that may participate in their regulation by ROS (see below), along with the fact that expression of several DSP is induced under various conditions of stresses.

3. Evidence of a Role for Reactive Oxygen Species in Map Kinase Activation Many studies have shown that bolus addition of exogenous tbCh, and exposure to radiation or to drugs such as menadione known to induce production of H2O2 lead to activation of the MAP kinases [26-29]. Modulation of GSH levels also plays a role in the activation of INK and p38 MAPK, as shown after treatment with alkylating agents [30]. ERK5 was discovered as a redox sensitive kinase [31] and is activated by HbOa in PC 12 cells [32]. Furthermore, many studies have implied involvement of ROS in MAP kinase activation after cell stimulation with various agents based on inhibition by catalase or compounds with antioxidants properties [33]. The prevention of HiCh accumulation by antioxidants blocked MAPK activation after stimulation by LPA, angiotensin and serotonin, all ligands for G-protein coupled receptors [34-36]. The small GTPase Rac, a component of the phagocyte NADPH oxidase that is ubiquitously expressed, appears to play a central role as a transducer of these receptor-activated redox signals [37, 38]. The mechanisms by which exogenous or endogenously produced ROS activate the MAP kinases are not well defined. Several mechanisms have been proposed for INK and p38 MAPK activation that involve ROS-dependent dissociation of a signalosome that maintains the pathway in an inactive state. ASK1, a MAP3K for JNK and p38 MAPK associates with reduced Trx in non-stressed cells. Oxidation of Trx by ROS activates ASK-1, leading to JNK activation [39], possibly through dimerization, as was reported after activation by ROSproduced TNFcc [40]. ASK1 knockout mice exhibited lower levels of JNK and p38MAPK activation in comparison to wild type after H2O2 or TNFa stimulation [41]. JNK associates with the monomeric form of glutathione S-transferase Pi (GSTp) and is inactive in non-stressed cells. JNK activation by ROS may result from oligomerization of GSTp and release from JNK [42]. Activation of the MAPKs could be the result of activation of the Src kinase family as cSrc activation is one of the earliest steps of the UV response leading to JNK activation [43]. cSrc was found to be required for the H2O2-induced activation of ERK5 [32;44] and JNK [45] while Fyn, another member of the Src family, appears to be involved in the H2O2-induced activation of ERK1/2, probably through activation of JAK2 upstream of Ras [46]. Nevertheless, the mechanisms for the redox sensitivity of these kinases are not known. The

310

M. Torres and H.J. Forman / Redox. Signaling and the Map Kinase Pathways

small GTPase Ras, which is an upstream activator of the ERKs and possibly other MAP kinases is directly altered by oxidation of a cysteine close to the guanine nucleotide binding site, resulting in activation [47, 48]. This study was among the first ones to suggest that alteration of essential thiols by ROS may be of significance for their role in activating signaling pathways. Further evidence came from studies with protein tyrosine phosphatases (PTPs) that contain in then* active site a cysteine (CX5R) that is critical for their activity and can be modified by oxidation. This cysteine is particularly reactive with HaCh because the surrounding residues in the active site allows for its ionization to a thiolate. Reaction of the thiolate with H2C>2 leads to formation of a sulfenic acid (-SOH) intermediate, which can quickly react with a thiol (-SH) such as GSH to produce a disulfide in the FTP, a form that is also catalytically inactive. The environment in the active site may also dictate the interaction with GSH. PTP1B, for example, contains many lysine residues in the active site cleft that could interact with glutathione while the active site of low molecular weight PTP contains aromatic residues. Data obtained in vivo and in vitro indicate that HaCh can specifically and reversibly inactivate PTPs [49, 50]. EGF stimulation was shown to reversibly inhibit PTP1B [51]. More recently, another study showed the transient and reversible inhibition of SHP-2 in response to PDGF that allowed activation of the ERKs [52], although the exact nature of the ROS-induced modification was not characterized. The DSPs that control the activity of the MAPKs also contain the cysteine signature motif. Expression of either wild type or catalytically inactive of the DSP, MKP-2, in endothelial cells resulted in reduction and enhancement, respectively, of the HaCh-induced activation of JNK [53], indicating that HaCh may alter the activity of the MAP kinases by modulating their activation by upstream activators or inactivation by phosphatases. Although MAPK activation by ROS produced endogenously has been associated with proliferation or apoptosis, only a few studies have so far demonstrated changes in gene expression as downstream events of their activation. In cardiac fibroblasts, angiotensin II (Angll), which induces the production of ROS, activated all three MAP kinases in an ROSdependent manner, resulting in increase IL-6 gene expression through phosphorylation of the CREB by the ERK and p38 MAPK pathway [54]. In cardiac endothelial cells, Angll induces expression of osteopontin through ERK activation in a ROS-dependent manner [55]. Expression of the early growth reponse-1 (Egr-1) transcription factor by cyclic strain is regulated by the ROS-mediated activation of the ERK pathway [56]. Endogenous production of ROS by Angll, PDGF and TNFa resulted in the MAP kinase-dependent increased expression of monocyte chemotactic protein MCP-1 [57]. Although still preliminary, the activation of the MAP kinases by ROS appears to affect gene regulation. 4. Erk Activation in Macrophages In phagocytes such as macrophages, superoxide production is regulated through receptor/ligand interaction during phagocytosis of bacteria or particles and upon stimulation with a variety of soluble agents and results in the assembly of the multicomponent NADPH oxidase through translocation of the p47phox/p67phfa/p4(fhox cytosolic proteins to the membrane and formation of a stable complex with the p22phox/gp9lphox transmembrane flavocytochrome [58]. This complex is in the right conformation for gp9lphox, the terminal oxidase, to transfer one electron to oxygen to form superoxide. Superoxide then quickly dismutates to hydrogen peroxide, which is highly diffusible. Homologs of gp91p*ax, referred to as Nox, have been identified in many nonphagocytic cells and may participate in agonist-regulated production of

f or res and H.J. Forman /Redox Signaling and the Map Kinase Pathwavs

evidence has indicated that in neutrophiis assembly of the oxidase occurs intracellularly, it is currently thought that assembly occurs at the plasma membrane in macrophages. Extensive literature has documented the role of these ROS in bacterial killing in neutrophiis where the azurophilic granule component, myeloperoxidase is essential for this process, although a recent paper challenges this concept [60]. In alveolar macrophages (AM), it is unclear whether ROS have a direct effect on bacterial killing. We propose mat the stimulated production of ROS by receptor interaction in AM plays a role in the activation of signaling pathways leading to expression of inflammatory genes. Our previous data showed that stimulation of rat AM with zymosan-activated serum (ZAS), a source of C5a that binds to a G-protein linked receptor, resulted in production of ROS, increase in the tyrosine phosphorylation of various proteins and activation of the ERKs and p38 MAPKs. Activation of the ERKs required the presence of HaOa, as extracellular catalase significantly reduced the activation of both ERK1 and ERK2 while marginally affecting that of p38 MAPKs [61]. The mechanisms by which HbO2 participate in the activation of the ERKs in rat AM is still unclear but are unlikely to be direct, as the activation of their upstream kinases, MEK1 & 2, was also inhibited by catalase [61]. Stimulation of the NADPH oxidase with ADP, which results in similar production of superoxide as that induced by ZAS, did not induce ERK activation, indicating that HiO2 alone does not appear to be sufficient for activation of the ERKs [62]. As previously mentioned, upstream activators of the ERKs could be involved in their ROS-regulated activation. This includes Ras, Src, Pyk2 or the Shc/Grb2 complex. We favor the hypothesis that inhibition of a protein tyrosine phosphatase by the burst-produced H^Oa allows downstream signaling to the ERKs (Figure 3). This is based on data recently obtained using a macrophage cell line, NR8383. We showed that pretreatment of the cells with vanadate, a well-known inhibitor of PTPs, allowed activation of the ERKs by ZAS in the presence of catalase, indicating that vanadate relieved the block by catalase and could substitute for HiOa [63]. We are currently investigating the catalase sensitivity of c-Raf and Ras to identify the potential target for the PTP in the ERK pathway, cRaf requires tyrosine phosphorylation for full activation and appears to be the most likely target, as it has also been shown to interact with PTP IB [64, 65]. Macrophages play an important role in inflammation as they produce ROS, ingest and kill bacteria and secrete various cytokines. Several studies have suggested that cytokine production by macrophages is a function of the redox status of the cells [66]. Modifying the GSH/GSSG ratio in human monocytes resulted in activation of p38 MAPK by LPS and increased production of IL-12 [67]. Production of TNFa induced in primary alveolar macrophages (AM) and in RAW 264.7 was inhibited by SOD and catalase [68]. Thus, ROSmediated activation of the MAP kinase pathways may play a significant role in regulation of the cytokine genes and may act in concert with the NF-KB pathway, a transcription factor that is also redox-sensitive and is critical for cytokine expression [69]. In Kupffer cells, the resident macrophage in liver, the production of TNFa was regulated by ROS-activated NF-KB pathway [70]. Thus, activation of the ERK pathway by ROS may play a significant role in the regulation of the inflammatory response. Future studies in our laboratory will address these issues in rat alveolar macrophages.

312

M. Torres and H.J. Forman / Redox Signaling and the Map Kinase Pathways

Rg. 3. Proposed model for the activation of the ERKs by ZAS In rat alveolar macrophages. H2O2 produced by the stimulation of the NADPH oxidase may inhibit a FTP that acts upon a component of the ERK signaling pathway, which must be tyrosine phosphorylated for downstream signaling.

5. Conclusions While a strong consensus exists that ROS and changes in the redox potential of the cells can affect the activation of the MAP kinases, further studies are required to determine the ROS targets in these pathways, which are more likely to be upstream of the MAP kinases and may differ with cell type and stimuli or with the site of ROS production. Particular attention may be given to proteins associated with the cascade components such as the GST proteins or to the role of the scaffold proteins. As the MAP kinases are critical regulators of transcription, it will be important to better understand how their activation by ROS regulate gene expression, in particular in the context of other signaling molecules that are also activated by ROS.

References 1- M. Sundaresan, Z-X. Yu, V. J. Ferraris, K. Irani and T. Finkel, Requirement for generation of HjO2 for plateletderived growth factor signal transduction, Science 270 (1995) 296-299. 2- Y. S. Bae, S. W. Kang, M. S. Seo, I. C. Baines, E. Tekle, P. B. Chock and S. G. Rhee, Epidermal growth factor (EGF)-induced generation of hydrogen peroxide. Role in EGF receptor-mediated tyrosine phosphoryiation, J. Biol. Chem. 272 (1997) 217-221.

M. Torres and H.J. Forman / Redox Signaling and the Map Kinase Pathways

313

3- V. J. Thannickal and B. L. Fanburg, Reactive oxygen species in cell signaling, Am. J Physiol Lung Cell Mol. Physiol 279 (2000) L1005-L1028. 4- H. J. Forman and M. Torres, Redox signaling in macrophages, Mol. Aspects Med. 22 (2001) 189-216. 5- W. Droge, Free radicals in the physiological control of cell function, Physiol Rev. 82 (2002) 47-95. 6- K. Hensley, K. A. Robinson, S. P. Gabbita, S. Salsman and R. A. Floyd, Reactive oxygen species, cell signaling, and cell injury, Free Radic. Biol. Med. 28 (2000) 1456-1462. 7- Kyriakis, J. M. and Avruch, J. Mammalian mitogen-activated protein kinase signal transduction pathways activated by stress and inflammation. Physiol.Rev. (2001) 81, 807-869. 8- G. Pearson, F. Robinson, G. T. Beers, B. E. Xu, M. Karandikar, K. Berman and M. H. Cobb, Mitogen-activated protein (MAP) kinase pathways: regulation and physiological functions, Endocr. Rev. 22 (2001) 153-183. 9- C. Peyssonnaux and A. Eychene, The Raf/MEK/ERK pathway: new concepts of activation, Biol. Cell 93 (2001) 53-62. 10- R. J. Davis, Signal transduction by the JNK group of MAP kinases, Cell 103 (2000) 239-252. 11- W. Kolch, To be or not to be: a question of B-Raf?, Trends Neurosci. 24 (2001) 498-500. 12- A. S. Dhillon, S. Meikle, Z. Yazici, M. Eulitz and W. Kolch, Regulation of Raf-1 activation and signalling by dephosphorylatton, EMBO J. 21 (2002) 64-71. 13- W. Kolch, Meaningful relationships: the regulation of the Ras/Raf/MEK/ERK pathway by protein interactions, Biochem. J. 351 (2000) 289-305. 14- M. J. Marinissen and J. S. Gutkind, G-protein-coupled receptors and signaling networks: emerging paradigms, Trends Pharmacol. Sci. 22 (2001) 368-376. 15- T. Wada, K. Nakagawa, T. Watanabe, G. Nishitai, J. Seo, H. Kishimoto, D. Kitagawa, T. Sasaki, J. M. Penninger, H. Nishina and T. Katada, Impaired synergistic activation of stress-activated protein kinase SAPK/JNK in mouse embryonic stem cells lacking SEK1/MKK4: different contribution of SEK2/MKK7 isoforms to the synergistic activation, J. Biol. Chem. 276 (2001) 30892-30897. 16- W. Sun, K. Kesavan, B. C. Schaefer, T. P. Garrington, M. Ware, N. L Johnson, E. W. Getfand and G. L. Johnson, MEKK2 associates with the adapter protein Lad/RIBP and regulates the MEK5-BMK1/ERK5 pathway, J. Biol. Chem. 276 (2001) 5093-5100. 17- H. J. Schaeffer, A. D. Catling, S. T. Eblen, L S. Collier, A. Krauss and M. J. Weber, MP1: A MEK binding partner that enhances enzymatic activation of the MAP kinase cascade, Science 281 (1998) 1668-1671. 18- S. J. Perry and R. J. Lefkowitz, Arresting developments in heptahelical receptor signaling and regulation, Trends Cell Biol. 12 (2002) 130-138. 19- J. Yasuda, A. J. Whit marsh, J. Cavanagh, M. Sharma and R. J. Davis, The JIP group of mitogen-activated protein kinase scaffold proteins, Mol. Cell Biol 19 (1999) 7245-7254. 20- K. Yamamoto, H. Ichijo and S. J. Korsmeyer, BCL-2 is phosphorylated and inactivated by an ASKI/Jun Nterminal protein kinase pathway normally activated at G2/M, Mol. Cell Biol 19 (1999) 8469-8478. 21- H. Enslen and R. J. Davis, Regulation of MAP kinases by docking domains, Biol. Cell 93 (2001) 5-14. 22- A. J. Whitmarsh and R. J. Davis, Transcription factor AP-1 regulation by mitogen-activated protein kinase signal transduction pathways, J Mol. Med 74 (1996) 589-607. 23- M. J. Marinissen, M. Chiariello, M. Pallante and J. S. Gutkind, A network of mitogen-activated protein kinases links G protein-coupled receptors to the c-jun promoter: a role for c-Jun NH2-terminal kinase, p38s, and extracellular signal-regulated kinase 5, Mol. Cell Biol. 19 (1999) 4289-4301. 24- M. Torres, F. L. Hall and K. A. O'Neill, Stimulation of human neutrophils with formyl-methionyl-leucylphenylalanine induces tyrosine phosphorylation and activation of two distinct mitogen-activated protein kinases, J. Immunol. 150 (1993) 1563-1576. 25- M. Camps, A. Nichols and S. Arkinstall, Dual specificity phosphatases: a gene family for control of MAP kinase function, FASEB J. 14 (2000) 6-16. 26- K. Z. Guyton, Y. S. Liu, M. Gorospe, Q. B. Xu and N. J. Holbrook, Activation of mitogen-activated protein kinase by H2O2 - Role in cell survival following oxidant injury, J. Biol. Chem. 271 (1996) 4138-4142.

314

M. Torres and H.J. Forman / Redox Signaling and the Map Kinase Pathways

27- Y. Y. C. Lo, J. M. S. Wong and T. F. Cruz, Reactive oxygen species mediate cytokine activation of c-Jun NH2terminal kinases, J. Biol. Chem. 271 (1996) 15703-15707. 28- X. T. Wang, J. L Martindate, Y. S. Liu and N. J. Holbrook, The cellular response to oxidative stress: influences of mitogen-activated protein kinase signalling pathways on cell survival, Biochem. J 333 (1998) 291-300. 29- H. Kamata and H. Hirata, Redox regulation of cellular signalling, Cell Signal. 11 (1999) 1-14. 30- D. Wilhelm, K. Bender, A. Knebel and P. Angel, The level of intracellular glutathione is a key regulator for the induction of stress-activated signal transduction pathways including Jun N-temninal protein kinases and p38 kinase by alkylating agents, Mol. Cell Biol. 17 (1997) 4792-4800. 31- J. Abe, M. Kusuhara, R. J. Ulevitch, B. C. Berk and J-D. Lee, Big mitogen-activated protein kinase 1 (BMK1) is a redox-sensitive kinase, J. Biol. Chem. 271 (1996) 16586-16590. 32- Y. Suzaki, M. Yoshizumi, S. Kagami, A. H. Koyama, Y. Taketani, H. Houchi, K. Tsuchiya, E. Takeda and T. Tamaki, Hydrogen Peroxide Stimulates c-Src-mediated Big Mitogen-activated Protein Kinase 1 (BMK1) and the MEF2C Signaling Pathway in PC12 Cells, potential role in cell survival following oxidative insults, J. Biol. Chem. 277(2002)9614-9621. 33- S. Hashimoto, Y. Gon, K. Matsumoto, I. Takeshita and T. Hone, N-acetytcysteine attenuates TNF-alphainduced p38 MAP kinase activation and p38 MAP kinase-mediated IL-8 production by human pulmonary vascular endothelial cells, Br. J. Pharmacol. 132 (2001) 270-276. 34- M. Sekharam, J. M. Cunnick and J. Wu, Involvement of lipoxygenase in lysophosphatidic acid-stimulated hydrogen peroxide release in human HaCaT keratinccytes, Biochem. J. 346 (2000) 751-758. 35- M. Kyaw, M. Yoshizumi, K. Tsuchiya, K. Kirima and T. Tamaki, Antioxidants inhibit JNK and p38 MAPK activation but not ERK 1/2 activation by angiotensin II in rat aortic smooth muscle cells, Hypertens. Res. 24 (2001)251-261. 36- E. L. Greene, O. Houghton, G. Collinsworth, M. N. Gamovskaya, T. Nagai, T. Sajjad, V. Bheemanathini, J. S. Grewal, R. V. Paul and J. R. Raymond, 5-HT(2A) receptors stimulate mitogen-activated protein kinase via H,O2 generation in rat renal mesangial cells, Am. J. Physiol Renal Physiol 278 (2000) F650-F658. 37- M. Sundaresan, Z. X. Yu, V. J. Ferrans, D. J. Sulciner, J. S. Gutkind, K. Irani, P. J. GoWschmidt-Clermont and T. Finkel, Regulation of reactive-oxygen-species generation in fibroblasts by Rac1, Biochem J. 318 (1996) 379382. 38- T. Joneson and D. Bar-Sagi, A Rac 1 effector site controlling mitogenesis through superoxide production, J. Biol. Chem. 273 (1998) 17991-17994. 39- M. Saitoh, H. Nishitoh, M. Fujii, K. Takeda, K. Tobiume, Y. Sawada, M. Kawabata, K. Miyazono and H. Ichijo. Mammalian thioredoxin is a direct inhibitor of apoptosis signal-regulating kinase (ASK) 1, EMBO J. 17 (1998) 2596-2606. 40- H. Liu, H. Nishitoh, H. Ichijo and J. M. Kyriakis, Activation of apoptosis signal-regulating kinase 1 (ASK1) by tumor necrosis factor receptor-associated factor 2 requires prior dissociation of the ASK1 inhibitor thioredoxin, Mol. Cell Biol. 20 (2000) 2198-2208. 41- K. Tobiume, A. Matsuzawa, T. Takahashi, H. Nishitoh, K. Morita, K. Takeda, O. Minowa, K. Miyazono, T. Noda and H. Ichijo, ASK1 is required for sustained activations of JNK/p38 MAP kinases and apoptosis, EMBO Rep. 2 (2001)222-228. 42- V. Adter, Z. M. Yin, S. Y. Fuchs, M. Benezra, L Rosario, K. D. Tew, M. R. Pincus, M. Sardana, C. J. Henderson, C. R. Wolf, R. J. Davis and Z. Ronai, Regulation of JNK signaling by GSTp, EMBO J. 18 (1999) 1321-1334. 43- Y. Devary, R. A. Gottlieb, T. Smeal and M. Karin, The mammalian ultraviolet response is triggered by activation of Src tyrosine kinases, Cell 71 (1992) 1081-1091. 44- J. Abe, M. Takahashi, M. Ishida, J. D. Lee and B. C. Berk, c-Src is required for oxidative stress-mediated activation of big mitogen-activated protein kinase 1 (BMK1), J. Biol. Chem. 272 (1997) 20389-20394. 45- M. Yoshizumi, J. Abe, J. Haendeter, Q. Huang and B. C. Berk, Src and Cas mediate JNK activation but not ERK112 and p38 kinases by reactive oxygen species, J Biol. Chem. 275 (2000) 11706-11712.

Torres and H.J. Forman / Redox Signaling and the Map Kinase Pathway*

46- J. Abe and B. C. Berk, Fyn and JAK2 mediate Ras activation by reactive oxygen species, J Biol. Chem. 274 (1999)21003-21010. 47- H. M. Lander, J. S. Ogiste, K. K. Teng and A. Novofrodsky, p21ras as a common signaling target of reactive free radicals and cellular redox stress, J. Biol. Chem. 270 (1995) 21195-21198. 48- H. M. Lander, An essential role for free radicals and derived species in signal transduction, FASEB J. 11 (1997) 118-124. 49- J. M. Denu and K. G. Tanner, Specific and reversible inactivation of protein tyrosine phosphatases by hydrogen peroxide: Evidence for a sulfenic acid intermediate and implications for redox regulation, Biochemistry 37 (1998)5633-5642. 50- J. R. Kim, H. W. Yoon, K. S. Kwon, S. R. Lee and S. G. Rhee, Identification of proteins containing cysteine residues that are sensitive to oxidation by hydrogen peroxide at neutral pH, Anal. Biochem. 283 (2000) 214221. 51- S. R. Lee, K. S. Kwon, S. R. Kim and S. G. Rhee, Reversible inactivation of protein-tyrosine phosphatase 1B in A431 cells stimulated with epidermal growth factor, J Biol. Chem. 273 (1998) 15366-15372. 52- T. C. Meng, T. Fukada and N. K. Tonks, Reversible oxidation and inactivation of protein tyrosine phosphatases in vivo, Mol. Cell 9 (2002) 387-399. 53- C. J. Robinson, C. M. Sloss and R. Plevin, Inactivation of JNK activity by mitogen-activated protein kinase phosphatase-2 in EAhy926 endothelial cells is dependent upon agonist- specific JNK translocation to the nucleus, Cell Signal. 13 (2001) 29-41. 54- M. Sano, K. Fukuda, T. Sato, H. Kawaguchi, M. Suematsu, S. Matsuda, S. Koyasu, H. Matsui, K. YamauchiTakihara, M. Harada, Y. Saito and S. Ogawa, ERK and p38 MARK, but not NF-kappaB, are critically involved in reactive oxygen species-mediated induction of IL-6 by angiotensin II in cardiac fibroblasts, Circ. Res. 89 (2001) 661-669. 55- Z. Xie, D. R. Pimental, S. Lohan, A. Vasertriger, C. Pligavko, W. S. Colucci and K. Singh, Regulation of angiotensin 11-stimulated osteopontin expression in cardiac microvascular endothelial cells: role of p42/p44 mitogen-activated protein kinase and reactive oxygen species, J. Cell. Physiol. 188 (2001) 132-138. 56- B. S. Wung, J. J. Cheng, Y. J. Chao, H. J. Hsieh and D. L. Wang, Modulation of Ras/Raf/extracellular signalregulated kinase pathway by reactive oxygen species is involved in cyclic strain-induced early growth response1 gene expression in endothelial cells, Circ. Res. 84 (1999) 804-812. 57- G. W. De Keulanaer, M. Ushio-Fukai, Y. Qiqin, A. B. Chung, R. P. Lyons, N. Ishizaka, K. Rengarajan, W. R. Taylor, R. W. Alexander and K. K. Griendling, Convergence of redox-sensitive and mitogen-activated protein kinase signaling pathways in tumor necrosis factor-a-mediated monocyte chemoattractant protein -1 induction in vascular smooth muscle cells, Arterioscler. Thromb. Vase. Biol 20 (2000) 385-391. 58- B. M. Babior, NADPH oxidase: An update, Blood 93 (1999) 1464-1476. 59- J. D. Lambeth, G. Cheng, R. S. Arnold and W. A. Edens, Novel homologs of gp91phox, Trends Biochem. Sci. 25 (2000) 459-461. 60- E. P. Reeves, H. Lu, H. L. Jacobs, C. G. Messina, S. Bolsover, G. Gabella, E. O. Potma, A. Warley, J. Roes and A. W. Segal, Killing activity of neutrophils is mediated through activation of proteases by K+ flux, Nature 416(2002)291-297. 61- M. Torres and H. J. Forman, Activation of several MAP kinases upon stimulation of rat alveolar macrophages: Role of the NADPH oxidase, Arch. Biochem. Biophys. 366 (1999) 231-239. 62- E. Gozal, H. J. Forman and M. Torres, ADP stimulates the respiratory burst without activation of ERK and AKT in rat alveolar macrophages, Free Radic. Biol. Med. 31 (2001) 679-687. 63- Torres, M. and Forman, H. J. Vanadate inhibition of protein tyrosine phosphatases mimics hydrogen peroxide in the activation of the ERK pathway in alveolar macrophages. Ann.NY Acad.Sci. In Press 64- J. R. Fabian, I. O. Daar and D. K. Morrison, Critical tyrosine residues regulate the enzymatic and biological activity of Raf-1 kinase, Mol. Cell Biol 13 (1993) 7170-7179. 65- T. Jelinek, P. Dent, T. W. Sturgill and M. J. Weber, Ras-induced activation of Raf-1 is dependent on tyrosine phosphorylation, Mol. Cell Biol. 16 (1996) 1027-1034.

316

M. Torres and H.J. Forman / Redox Signaling and the Map Kinase Pathways

66- Y. Murata, T. Shimamura and J. Hamuro, The polarization of T(h)1/T(h)2 balance is dependent on the intracellular thiol redox status of macrophages due to the distinctive cytokine production, Int. Immunol. 14 (2002)201-212. 67- M. Utsugi, K. Dobashi, Y. Koga, Y. Shimizu, T. Ishizuka, K. lizuka, J. Hamuro, T. Nakazawa and M. Mori, Glutathkme redox regulates lipopolysaccharide-induced IL-12 production through p38 mttogen-activated protein kinase activation in human monocytes: role of glutathione redox in IFN-gamma priming of IL-12 production, J. Leukoc. Biol. 71 (2002) 339-347. 68- Y. Rojanasakul, J. Ye, F. Chen, L Wang, N. Cheng, V. Castranova, V. Valryathan and X. Shi, Dependence of NF-kappaB activation and free radical generation on silica-induced TNF-alpha production in macrophages, Mol. Cell Biochem. 200 (1999) 119-125. 69- Y. M. Janssen-Heininger, M. E. Poynter and P. A. Baeuerie, Recent advances towards understanding redox mechanisms in the activation of nuclear factor kappaB, Free Radic. Biol Med. 28 (2000) 1317-1327. 70- M. L. Rose, I. Rusyn, H. K. Bojes, J. Belyea, R. C. Catttey and R. G. Thurman, Role of Kupffer cells and oxidants in signaling peroxisome proliferator-induced hepatocyte proliferation, Mutat. Res. 448 (2000) 179-192.

Thiol Metabolism and Redox Regulation of Cellular Functions A. Pompella etal. (Eds.) IOS Press, 2002

31'

Redox Regulation of Mitochondrial Permeability Transition: Contrasting Effects of Lipoic Acid and Its Positively Charged Analog LA-plus Oren TIROSH1*, Shani SfflLO1, Anna ARONIS1 and Chandan K. SEN2 lnst. of Biochemistry, food Science and Nutrition, The Hebrew University of Jerusalem, Rehovot 76100, Israel; 2Lab. of Molecular Medicine, Dept. of Surgery, Davis Heart & Lung Research Institute, The Ohio State University Medical Center, Columbus, USA

l

I, Introduction /. 1. Mitochondria andROS Oxidative phosphorylation is arguably the most important energy transudation process in eukaryotes. In eukaryotes, oxidative phosphorylation occurs in the mitochondria. The chemiosmotic theory of mitochondria! energy production suggests the formation of a proton-motive force that represents the capacity to generate energy-rich ATP molecules (Mitchell, 1979; Mitchell and Moyle, 1967). Over 90% of the oxygen consumption in vivo has been attributed to mitochondrial respiration (Babcock and Wikstrom, 1992). Mitochondria are considered the main source of ROS in eukaryotic cells (Boveris and Chance, 1973; Boveris et al., 1972; Cadenas and Davies, 2000; Chance et al., 1979). The realization that energy consumption by mitochondria can generate oxygen radicals has linked the free radical theory of aging to mitochondrial functionality (Harman, 1972; Miquel et al., 1992). Mitochondria from post-mitotic cells use Oa at a high rate and may release oxygen radicals that overwhelm cellular antioxidant defenses (Sastre et al., 2000). Indeed, mitochondria are the major source of superoxide anion production in cells. During the transfer of electrons to molecular oxygen, an estimated 1 to 5% of electrons in the respiratory chain "leak" to form superoxide radicals (Boveris and Chance, 1973; Chance et al., 1979; Sastre et al., 2000). 1.2. Cell Death The last decade has witnessed a major improvement of our understanding of the fundamental mechanisms that lead to cell death (Agarwal et al., 1998; Gottlieb and Oren, 1998; Nagata, 1997; Oren, 1999; Steller, 1995; Wallach, 1997). Efforts have been made to classify cellular death into two major categories: necrosis and apoptosis or programmed cell death (Hampton and Orrenius, 1997; Nicotera et al., 1997; Samali et al., 1999). It should be noted, however, that death processes sharing the characteristics of apoptosis and necrosis has been evident as well (Tirosh et al., 2000). Somatic cells are capable of self-destruction by activating an intrinsic death program, which is usually turned on when cells are no longer needed or have become seriously damaged. This biological death cascade, which has been termed apoptosis, is often associated with specific morphological and biochemical

318

O. Tirosh et at. / Redox Regulation ofMitochondrial Permeability Transition

characteristics (Steller, 1995). Recently it has been suggested that mitochondria is involved in the redox regulation of cell death and aging processes (Zhang and Herman, 2002). 1.3. Mitochondnal Control of Cell Death One of the hallmarks of apoptosis is the early and temporal dissipation of the mhochondrial membrane potential (Kroemer et al., 1998; Kroemer et al., 1997; Petit et al., 1995; Petit et al., 1996). Several mechanisms are known by which mitochondria play a role in cell death, and their effects may be inter-related (Blanc et al., 2000; Green and Reed, 1998; Korsmeyer et al., 2000; Kowaltowski et al., 1996b; Kowaltowski et al., 2001). These mechanisms are: i) disruption of electron transport, oxidative phosphorylation, and ATP production. The cellular level of ATP has been shown to determine the fate of the apoptotic process (Richter et al., 1996); ii) release of proteins that trigger activation of caspase-family proteases (cytochrome c, AEF) which in turn are responsible for the execution of the apoptotic process. Mitochondria are known to serve as a pool for factors that can initiate and exacerbate the apoptosis; and iii) alteration of cellular reduction-oxidation (redox) potential because of excess production of ROS (Li et al., 1997; Li et al., 1998; Murphy et al., 1989; Tan et al., 1998; Tirosh et al., 1999; Tirosh et al., 2000). MPT is triggered by an increase in the inner mitochondria! membrane's permeability most easily observed after matrix Ca2+ accumulation (Gunter and Pfeiffer, 1990). Although MPT can be favored by a large series of heterogeneous compounds and conditions, it is generally agreed that it is mediated by the opening of a cyclosporin A (CsA)-sensitive complex channel (Scorrano et al., 1997; Zoratti and Szabo, 1995).Evidence that pore opening is voltage-controlled in intact isolated mitochondria is largely based on the effects of the protonophoric uncoupler carbonyl cyanide p-(trifluoromethoxy)phenyl-hydrazone (FCCP), whereas other uncouplers have not been reported to induce MPT (Bernard! and Petronilli, 1996). ROS (e.g., t-butyl hydroperoxide, peroxynitrite) have been suggested to facilitate the process of MPT pore opening (Costantini et al., 2000; Costantini et al., 1996; Kowaltowski et al., 2001; Zago et al., 2000). The mechanism of action has been suggested to involve cross-linking of critical thiols in the MPT pore region. In addition, elevation in the production of endogenous mitochondrial ROS was demonstrated in Pi-induced MPT (Kowaltowski et al., 1996a). However, it is still not clear whether ROS are the effectors of MPT or are simply the released product of several PT inducers. In a recent publication, induction of MPT was shown not to be facilitated by ROS as demonstrated by several antioxidants, which were unable to prevent MPT (Qu et al., 2001).

FCCPIOuM

FCCP 0.

10 min

^

—

^_ No FCCP

Fig. 1. Protective effect of FCCP against oxidant (t-butyl hydroperoxide; TBH) challenge of rat liver mitochondria. Swelling effect: FCCP was added after exposure to 250 ^M TBH.

O. Tirosh c{ al. / Redox Regulation of Mitochondrial Permeability Transition

319

2. Mitochondrial Depolarization and Oxidative Stress Exposure of rat liver mitochondria loaded with calcium to 250 ^M of t-butyl hydroperoxide promoted extensive swelling (Figure 1). FCCP added to the rat liver mitochondria after exposure to the oxidant prevented mitochondrial swelling in a dose-dependent fashion. Depolarization of the mitochondrial membrane prevented oxidant-induced swelling by closing the MPT-pore as was evaluated by PEG500 contraction assay. FCCP afforded protection when added after exposure to the oxidant. Therefore, attenuated membrane potential under oxidative stress may permit the mitochondria to regulate swelling by releasing calcium. Introducing a small amount of FCCP together with an oxidant exacerbated calcium release. This observation implies that mitochondria exposed to oxidative stress are more receptive to low membrane potential-dependent calcium release. It is therefore plausible that the level of mitochondrial polarization governs the thiol oxidation status of the MPT pore. In this way, the state of mitochondrial polarization may regulate MPT pore opening.

intracellular

extracellular

LA

Mitochondrial Lipoamide dehydrogenase

L

r

cystine

LDH

(NApH*NAD + ) GR, FR (NADPH* NADP+) /

/ yGCS cysteine

DHLA

+

•

cysteine

••••••••^^^ Q§H J glutathione synthetase

DHLA

Fig. 2. Cellular pathways for the conversion of atpha-Hpoic acid (LA) to dihydrolipoic acid (DHLA) and lipoate mediated upregulation of cellular glutathione (6SH) biosynthesis via increase in cysteine bioavailability. TR, thioredoxin reductase; GR, glutathione reductase; yGCS, gamma glutamyl cysteine synthetase; LDH, lipoamide dehydrogenase.

320

O. Tirosh et al. / Redox Regulation of Mitochondrial Permeability Transition

3. Regulation of Mitochondrial Permeability Transition Exposing mitochondria to various thiol-containing compounds has been suggested to facilitate high amplitude swelling (Schweizer and Richter, 1996) of isolated mitochondria. On the other hand, thiol-antioxidants were reported to prevent Pi induced MPT pore opening and to protect against high amplitude swelling (Kowaltowski et al., 1996b; Kowaltowski et al., 1998). It is therefore important to investigate the specific effect of each thiol compound and to elucidate whether or not it is capable of affecting mitochondria! function. R-a-Lipoic acid (R-LA) is a naturally occurring compound present as a co-factor in a number of mitochondrial enzymes that are involved in metabolism and energy production. ct-Lipoic acid (LA) was first isolated in 1951 by Reed and colleagues. LA is an eight-carbon compound containing two sulfur atoms in a dithiolane ring structure. The two enantiomers of LA are the S form and the R form (Figure 2). In its free form, LA is considered to be a powerful antioxidant, functioning as a reactive oxygen species scavenger (Packer, 1994; Packer, 1998; Packer et al., 1997; Packer et al., 1995). Previously it has been demonstrated that active (reduced) a-dihydrolipoic acid (DHLA) can potentiate apoptosis in Jurkat T-cells treated with anti CD95 antibodies (Sen et al., 1999). The intracellular events, which were potentiated by LA in these cells undergoing apoptosis, were increased cytosolic calcium and loss of mitochondrial membrane potential. We developed a water soluble positively charged (at pH 7.4) analogue of lipoamide, LA-plus (Figure 3) (Sen et al., 1998). This two sulfur-bearing antioxidant has been investigated in the context of affecting MPT.

A l p h a - L i p o i c acid

L A - P l u s ; 5-[l,2] Dithiolan-3-yl-pentanoic acid (2-dimethylam ino-ethyl)-amide

Fig. 3. Chemical structures of lipoic acid and LA-plus

O, Tirosh et al. / Redox Regulation of Mitochondrial Permeability Transition

321

3.1. Uptake of LA-plus and LA into Mitochondria Exposure of isolated liver or brain mitochondria to 150 uM LA-plus or LA resulted in more of the total amount (sum of oxidized and reduced forms) of LA-plus relative to LA being taken up. More of the reduced form of LA-plus (DHLA-plus) than DHLA was found in liver as well as brain mitochondria indicating that reduction of LA-plus is far more efficient than LA (Figure 4). This is consistent with the enzyme kinetics that we (Tirosh et al., 1999) and others (Korotchkina et al., 2001) have reported indicating that the LA-plus analogue is reduced by dihydrolipoamide dehydrogenase far better than LA. Low amounts of DHLAplus have been detected in the mitochondrial suspension medium. This suggests that some DHLA-plus is release from the mitochondria. It was clear that LA-plus is more efficiently taken up and reduced in liver mitochondria than LA. LA-plus therefore is better suited than LA to serve as a reductant in mitochondria.

amount (nmol)

Liver Mitochondria

Liver Medium

Fig. 4. LA or LA-plus uptake and reduction by rat brain mitochondria and corresponding extra-mitochondrial content. Mitochondria (2 mg) were treated with 150 \M of the compounds for 10 min. Mitochondria were separated from the medium and LA or LA-plus and their corresponding reduced forms were analyzed as described in Tirosh et al 1999: LA-plus, closed bars; DHLA-plus, open bars; LA, gray cross hatched bars; DHLA, vertical line bars.

3.2. Mitochondrial Membrane Potential, Swelling and Calcium Cycling Mitochondria have been used in flow cytometry and their size has been determined. By diluting the mitochondria 500 times we find that preloaded rhodamine 123 leaves the mitochondria upon loss of membrane potential thus allowing the analysis of mitochondrial membrane potential. Using this approach it was strikingly found that free LA at 150 uM (not the protein bound LA) alone induced a loss of mitochondrial membrane potential, whereas LA-plus at 150 uM actually protected and maintained mitochondrial membrane potential over time (Figure 5). The balance of reduced/oxidized forms of the compounds indicate that a high ratio of disulfide (S-S) to dithiols (-SH) can trigger mitochondrial permeability transition while changing the ratio to low S-S and high SH can actually stabilize the mitochondria against the induction of MPT. The fact that LA or LA-plus are exogenous compounds and not part of the total thiol pool of the mitochondria is important. Control of MPT induction by

322

O. Tirosh el al. / Redox Regulation of Mitochondrial Permeability Transition

exogenous thiol status can be explained by a quick equilibrium between the thiols that control the formation of the MPT pore and the thiols that are being delivered to the mitochondria. In spectrophotometer monitored swelling of isolated liver and brain mitochondria it was observed that LA alone at 150 uM or in the presence or absence of oxidants, induces swelling of mitochondria, whereas LA-plus at 150 uM maintains the integrity of mitochondria whether or not exposed to oxidants. Brain mitochondria were found to be extremely sensitive to LA (150 uM) induced swelling but not to TBH treatment. Therefore, the effect of LA was most probably not facilitated by an oxidationrelated mechanism, but due to a specific disulfide interaction. Both stereoisomers of LA induced swelling in rat liver mitochondria suggesting that the disulfide bond was the mediator of the MTP promoting effect. Time control (min) i

10

R-Lipoate R-LA-plus TBH

_1

20

30

.i Rhodamine 123

Fig. 5. Changes in mitochondria! membrane potential of succinate energized rat liver mitochondria. Isolated rat liver mitochondria (2 mg/ ml) in MSH were treated with rotenone 2 \>M, calcium 60 nmol/ mg protein, succinate 5 mM and LA, LA-plus (150 uM) or TBH (250 \AK). Rh123, 2 ng/ml was added. At the different time intervals indicated, samples were diluted 500 times (2 nJ in 1 ml MSH buffer) and analyzed by flow cytometry immediately following dilution. Histograms represent 10,000 events. Histograms are presented as Rh123 fluorescence in log scale (x-axis) Vs* events (Yaxis). Shifts of points in the histogram to the left indicate loss of Rh123 fluorescence and therefore a decline of mitochondrial membrane potential.

Previously, two independent groups (Jayanthi et al., 1991; Schweizer and Richter, 1996) have shown Ca2+ release induced by LA in isolated liver mitochondria. Our own observations confirm those findings. However, LA-plus does not induce Ca2* release, at concentrations below 220 uM. At 75 to 150 uM, LA-plus did not trigger Ca2+ release. Under conditions of oxidant stress induced by TBH, mitochondria released Ca2* in the presence of LA (normal biphasic release) faster than with TBH alone. However, in the presence of LA-plus there is a very long lag phase in the TBH induced Ca2+ release. Furthermore when the Ca2+ release is initiated it is only a slow linear release. This can be interpreted as an inhibitory affect of LA-plus against TBH induced Ca2+ release. This is contrary to the effect of LA at 150 uM, which destabilizes mitochondria and potentates the

O. Tirosh et al. / Redox Regulation ofMitochondrial Permeability Transition

323

release of calcium. In view of the swelling and membrane potential results its likely that high levels of oxidized form LA is inducing extensive Ca2+ cycling and this is eventually causing loss of membrane potential and swelling. It is clear from the uptake experiments that more LA-plus is taken up and reduced by the mitochondria than LA. The enzyme kinetic experiment supports this observation (Tirosh et al., 1999). Thus it is highly likely that the reduced form of LA-plus could be keeping the vicinal thiol reduced, thus inhibiting Ca2+ release. On the other hand when the oxidized form is present in higher amounts (as in the case of LA) there is Ca2+ release due to oxidation of the vicinal thiol. Thus, it may be predicted that LA would facilitate cell death and this is exactly what has been observed in Jurkat cells (Sen et al., 1999). Such toxic effect of LA is possibly due in part to an interaction and destabilization of cell mitochondria, which may explain the synergistic potentiation in killing tumor cells exposed to doxorubicine (Dovinova et al., 1999)

4. Selenium-Thiol Interaction and Mitochondrial Permeability Transition Sodium selenite is a common dietary form of selenium, recognized as essential in animal and human nutrition (Combs, 1999; Combs and Gray, 1998). In the amino acid bound form, selenocysteine, it is a component of a number of antioxidant enzymes, e.g. the enzyme glutathione peroxidase and thioredoxin reductase (Arner and Holmgren, 2000; BrigeliusFlohe et al., 2000). Selenium supplementation induces immune boosting, chemo-protective as well as anticancer activities. Such activities have been associated with selenium intake that corrects for nutritionally deficient status in animals. A higher intake of selenium in mice, 2 ppm, prevented mammary tumorigenesis more effectively, and independently of glutathione peroxidase levels of expression (Medina et al., 1983). Therefore, it seems that selenium intake at concentration higher than those associated with maximal expression of the selenocysteine-containing enzymes is beneficial (Combs, 1999; Combs and Gray, 1998). Little information is available on the biological activity of selenium or on its function in its enzyme-free form; most experiments on the topic have involved its activity while incorporated into selenoproteins (Arner and Holmgren, 2000; Molina and Garcia, 1997; Schulz et al., 1999). Induction of apoptosis of cancer cells is the preferred way of eliminating them. In addition, an efficient and functional apoptotic process in normal cells prevents malignant transformation and helps multicellular organisms with developmental processes. There is a known connection between selenium and apoptosis (Ganther, 1999; Ip and Dong, 2001; Shen et al., 2000; Wei et al., 2001; Yang et al., 2000). Selenium may facilitate the reactions of cysteine residues by transient formation of more reactive S-Se intermediates leading to cell death (Ganther, 1999). In recent years, efforts have been made to explain the pro-apoptotic effect of selenium. It was shown in a human hepatic cell line and human hepatoma cell line that SeOa prompts apoptosis in correlation with downregulation of Bcl-2 and up-regulation of p53 levels (Ip and Dong, 2001; Wei et al., 2001). However, selenium can trigger apoptosis independent of DNA damage in cells having a null p53 phenotype (Ganther, 1999). It has been suggested that the cell-cycle protein kinase cdk2 and protein kinase C are strongly inhibited by various forms of selenium due to the formation of selenium adducts of the selenotrisulfide (S-Se-S) or selenenylsulfide (S-Se) type, or catalysis of disulfide formation (Ganther, 1999). Ebselen, a selenium-containing compound, has been recently found to induce apoptosis via induction of mitochondrial permeability transition (MPT) (Zhang and Herman, 2002). Interaction with thiols is a major aspect of selenium biochemistry. The reaction of selenite with the reduced form of glutathione (GSH) leads to the formation of ROS (Shen et

O. Tirosh el al. / Redox Regulation of Mitochondrial Permeability Transition

324

al., 2000). Such oxidizing activity may regulate the opening of the MPT pore. However, monitoring of ROS production levels in the mitochondria! matrix showed no elevation in mitochondria-derived ROS. Moreover, thiol antioxidants not only failed to prevent the effect of selenium, thiols potentiated Se-induced opening of the MPT pore (Figure 6). Therefore, it is possible that the reducing power of mitochondrial thiols facilitated rapid sodium selenite reduction leading to selenium-dependent MPT pore opening, swelling and cytochrome c release. It is therefore suggested that selenium in the form of sodium selenite interacts with thiols. Such interaction facilitates MPT pore opening and supports apoptosis. Thus, interaction between two antioxidants, thiols and selenium, may lead to a ROS-independent pro-apoptotic effect.

0.5 OD

NACSOuM SelOuM 54TJmi SelOuM+NACSOuM

Fig. 6. Effect of selenium and thiols on MPT pore opening. Swelling monitored at CDs*) in succinate-energized mitochondria in the presence of rotenone: 10 \M sodium selenite and 50 \M MAC were used to induce swelling. Adding thiols to selenium caused intensified high-amplitude swelling.

In conclusion, works reviewed in this chapter suggest that low mitochondrial membrane potential may serve as a protective barrier against ROS-induced mitochondria! degeneration and that thiol antioxidants such as lipoic acid may support the apoptotic processes via its effects on the mitochondria. Oxidized (disulfide) thiols such as lipoic acid directly opens the MPT pore while reduced thiols need an electron transfer intermediate molecule such as selenium to facilitate MPT pore opening. On the other hand excessive reduced thiol (SH) accumulation in the mitochondrial matrix such as in the case of LA-plus protect and prevent MPT. LA-plus has outstanding therapeutic potential in disorders related to mitochondriopathy.

References Agaiwal, M. L, Taylor, W. R., Chernov, M. V., Chemova, O. B.. and Stark, G. R. (1998). The p53 network, J Biol Chem 273,1-4. Amer, E. S., and Holmgren, A. (2000). Physiological functions of thioredoxin and thioredoxin reductase, Eur J Biochem 267,6102-9. Babcock, G. T., and Wikstrom, M. (1992). Oxygen activation and the conservation of energy in cell respiration, Nature 356, 301-9. Bemardi, P., and Petronilli, V. (1996). The permeability transition pore as a mitochondrial calcium release channel: a critical appraisal, J Bioenerg Biomembr 28,131-8.

O. Tirosh et al. / Redox Regulation of Mitochondrial Permeability Transition

325

Blanc, C., Deveraux, Q. L., Krajewski, S., Janicke, R. U., Porter, A. G., Reed, J. C., Jaggi, R., and Marti, A. (2000). Caspase-3 is essential for procaspase-9 processing and cisplatin-induced apoptosis of MCF-7 breast cancer cells, Cancer Res 60, 4386-90. Boveris, A., and Chance, B. (1973). The mitochondria! generation of hydrogen peroxide. General properties and effect of hyperbaric oxygen, Biochem J 134, 707-16. Boveris, A., Oshino, N., and Chance, B. (1972). The cellular production of hydrogen peroxide, Biochem J 128, 617-30. Brigelius-Flohe, R., Maurer, S., Lotzer, K., Bol, G., Kallionpaa, H., Lehtolainen, P., Viita, H., and Yla-Herttuala, S. (2000). Overexpression of PHGPx inhibits hydroperoxide-induced oxidation, NFkappaB activation and apoptosis and affects oxLDL-mediated proliferation of rabbit aortic smooth muscle cells, Atherosclerosis 152, 307-16. Cadenas, E., and Davies, K. J. (2000). Mitochondrial free radical generation, oxidative stress, and aging, Free Radic Biol Med 29, 222-30. Chance, B., Sies, H., and Boveris, A. (1979). Hydroperoxides metabolism in mammalian organs, Physiological Reviews 59, 527-605. Combs, G. F., Jr. (1999). Chemopreventive mechanisms of selenium, Med Klin 94 Suppl3,18-24. Combs, G. F., Jr., and Gray, W. P. (1998). Chemopreventive agents: selenium, Pharmacol Ther 79, 179-92. Costantini, P., Belzacq, A. S., Vieira, H. L, Larochette, N., de Pablo, M. A., Zamzami, N., Susin, S. A., Brenner, C., and Kroemer, G. (2000). Oxidation of a critical thiol residue of the adenine nucleotide translocator enforces Bcl-2-independent permeability transition pore opening and apoptosis, Oncogene 19, 307-14. Costantini, P., Chemyak, B. V., Petronilli, V., and Bernard!, P. (1996). Modulation of the mitochondria! permeability transition pore by pyridine nucleotides and dithiol oxidation at two separate sites, J Biol Chem 271,6746-51. Dovinova, I., Novotny, L., Rauko, P., and Kvasnicka, P. (1999). Combined effect of lipoic acid and doxorubicin in murine leukemia, Neoplasma 46,237-41. Ganther, H. E. (1999). Selenium metabolism, selenoproteins and mechanisms of cancer prevention: complexities with thioredoxin reductase, Carcinogenesis 20,1657-66. Gottlieb, T. M., and Oren, M. (1998). p53 and apoptosis, Semin Cancer Biol 8, 359-68. Green, D. R., and Reed, J. C. (1998). Mitochondria and apoptosis, Science 281,1309-12. Gunter, T. E., and Pfeiffer, D. R. (1990). Mechanisms by which mitochondria transport calcium, Am J Physiol 258, C755-86. Hampton, M. B., and Orrenius, S. (1997). Dual regulation of caspase activity by hydrogen peroxide: implications for apoptosis, FEBS Lett 474, 552-6. Harman, D. (1972). The biologic dock: the mitochondria?, J Am Geriatr Soc 20,145-7. Ip, C., and Dong, Y. (2001). Methylselenocysteine modulates proliferation and apoptosis biomarkers in premalignant lesions of the rat mammary gland, Anticancer Res 21, 863-7. Jayanthi, S., Jayanthi, G., and Varalakshmi, P. (1991). Effect of DL alpha-lipoic acid on some carbohydrate metabolising enzymes in stone forming rats, Biochem Int 25,123-36. Korotchkina, L. G., Yang, H., Tirosh, O., Packer, L., and Patel, M. S. (2001). Protection by thiols of the mitochondria! complexes from 4-hydroxy-2- nonenal, Free Radic Biol Med 30, 992-9. Korsmeyer, S. J., Wei, M. C., Saito, M., Weiler, S., Oh, K. J., and Schlesinger, P. H. (2000). Proapoptotic cascade activates BID, which oligomerizes BAK or BAX into pores that result in the release of cytochrome c, Cell Death Differ 7,1166-73. Kowaltowski, A. J., Castilho, R. F., Grijalba, M. T., Bechara, E. J., and Vercesi, A. E. (1996a). Effect of inorganic phosphate concentration on the nature of inner mitochondria! membrane alterations mediated by Ca2+ ions. A proposed model for phosphate-stimulated lipid peroxidation, J Biol Chem 27f, 2929-34. Kowaltowski, A. J., Castilho, R. F., and Vercesi, A. E. (1996b). Opening of the mitochondria! permeability transition pore by uncoupling or inorganic phosphate in the presence of Ca2+ is dependent on mitochondrial-generated reactive oxygen species, FEBS Lett 378,150-2. Kowattowski, A. J., Castilho, R. F., and Vercesi, A. E. (2001). Mitochondrial permeability transition and oxidative stress, FEBS Lett 495,12-5. Kowaltowski, A. J., Netto, L. E., and Vercesi, A. E. (1998). The thiol-specific antioxidant enzyme prevents mitochondria! permeability transition. Evidence for the participation of reactive oxygen species in this mechanism, J Biol Chem 273,12766-9. Kroemer, G., Dallaporta, B., and Resche-Rigon, M. (1998). The mitochondria! death/life regulator in apoptosis and necrosis, Annu Rev Physiol 60,619-42. Kroemer, G., Zamzami, N., and Susin, S. A. (1997). Mitochondrial control of apoptosis, Immunol Today 18, 44-51.

326

O. Tirosh et al. / Redox Regulation of Mitochondrial Permeability Transition

Li, Y., Maher, P., and Schubert, D. (1997). A role for 12-lipoxygenase in nerve cell death caused by glutathione depletion, Neuron 19,453-63. Li, Y., Maher, P., and Schubert, D. (1998). PhosphatkJylcholine-specrfic phospholipase C regulates glutamate- induced nerve cell death, Proc Natl Acad Sci U S A 95,7748-53. Medina, D., Lane, H. W., and Tracey, C. M. (1983). Selenium and mouse mammary tumorigenesis: an investigation of possible mechanisms, Cancer Res 43,2460s-2464s. Miquel, J., de Juan, E., and Sevila, I. (1992). Oxygen-induced mitochondria! damage and aging, Exs 62. 47-57. Mitchell, P. (1979). Keilin's respiratory chain concept and its chemiosmotic consequences, Science 206,1148-59. Mitchell, P., and Moyle, J. (1967). Chemiosmotic hypothesis of oxidative phosphorylation, Nature 213,137-9. Molina, H., and Garcia, M. (1997). Enzymatic defenses of the rat heart against lipid peroxidation, Mech Ageing Dev 97,1-7. Murphy, T. H., Miyamoto, M., Sastre, A., Schnaar, R. L., and Coyle, J. T. (1989). Glutamate toxicity in a neuronal cell line involves inhibition of cystine transport leading to oxidative stress, Neuron 2,1547-58. Nagata, S. (1997). Apoptosis by death factor. Cell 88, 355-65. Nicotera, P., Ankarcrona, M., Bonfoco, E., Orrenius, S., and Upton, S. A. (1997). Neuronal necrosis and apoptosis: two distinct events induced by exposure to glutamate or oxkJative stress, Adv Neural 72, 95-101. Oren, M. (1999). Regulation of the p53 tumor suppressor protein, J Biol Chem 274, 36031-4. Packer, L. (1994). Antioxidant properties of lipoic add and its therapeutic effects in prevention of diabetes complications and cataracts, Ann N Y Acad Sci 738, 257-64. Packer, L. (1998). alpha-Lipoic acid: a metabolic antioxklant which regulates NF-kappa B signal transduction and protects against oxidative injury, Drug Metab Rev 30,245-75. Packer, L., Roy, S., and Sen, C. K. (1997). Alpha-lipoic acid: a metabolic antioxklant and potential redox modulator of transcription, Adv Pharmacol 38, 79-101. Packer, L., Witt, E. H., and Tritschler, H. J. (1995). alpha-Lipoic acid as a biological antioxidant, Free Radic Biol Med 19,227-50. Petit, P. X., Lecoeur, H., Zom, E., Dauguet, C., Mignotte, B., and Gougeon, M. L. (1995). Alterations in mitochondrial structure and function are early events of dexamethasone-induced thymocyte apoptosis, J Cell Biol 130,157-67. Petit, P. X., Susin, S. A., Zamzami, N., Mignotte, B., and Kroemer, G. (1996). Mitochondria and programmed cell death: back to the future, FEBS Lett 396,7-13. Qu, B., Li, Q., Wong, K. P., Tan, T. M., and Halliwell, B. (2001). Mechanism of dofibrate hepatotoxicity: mitochondrial damage and oxidative stress in hepatocytes, Free Radic Biol Med 31, 659-69. Richter, C., Schweizer, M., Cossarizza, A., and Franceschi, C. (1996). Control of apoptosis by the cellular ATP level, FEBS Lett 378,107-10. Samali, A., Nordgren, H., Zhivotovsky, B., Peterson, E., and Orrenius, S. (1999). A comparative study of apoptosis and necrosis in HepG2 cells: oxjdant-induced caspase inactivation leads to necrosis, Biochem Biophys Res Commun 255,6-11. Sastre, J., Pallardo, F. V., and Vina, J. (2000). Mitochondrial oxidative stress plays a key role in aging and apoptosis [In Process Citation], IUBMB Life 49,427-35. Schulz, H. U., Niederau, C., Klonowski-Stumpe, H., Halangk, W., Luthen, R., and LJppert, H. (1999). Oxidative stress in acute pancreatitis, Hepatogastroenterology 46,2736-50. Schweizer, M., and Richter, C. (1996). Stimulation of Ca2+ release from rat liver mitochondria by the dithiol reagent alpha-lipoic add, Biochem Pharmacol 52,1815-20. Scorrano, L., Petronilli, V., and Bemardi, P. (1997). On the voltage dependence of the mitochondrial permeability transition pore. A critical appraisal, J Biol Chem 272,12295-9. Sen, C. K., Sashwati, R., and Packer, L. (1999). Fas mediated apoptosis of human Jurkat T-cells: intracellular events and potentiation by redox-active alpha-lipoic acid, Cell Death Differ 6, 48191. Sen, C. K., Tirosh, O., Roy, S., Kobayashi, M. S., and Packer, L. (1998). A positively charged alphalipoic acid analogue with increased cellular uptake and more potent immunomodulatory activity, Biochem Biophys Res Commun 247,223-8. Shen, H., Yang, C., Liu, J., and Ong, C. (2000). Dual role of glutathione in selenite-induced oxidative stress and apoptosis in human hepatoma cells, Free Radic Biol Med 28,1115-24. Steller, H. (1995). Mechanisms and genes of cellular suicide, Science 267,1445-9. Tan, S., Sagara, Y., Liu, Y., Maher, P., and Schubert, D. (1998). The regulation of reactive oxygen species production during programmed cell death, J Cell Biol 141,1423-32.

O. Tirosh el al. / Redox Regulation of Mitochondria! Permeability Transition

Tirosh, O., Sen, C. K., Roy, S., Kobayashi, M. S., and Packer, L (1999). Neuroprotective effects of alpha-lipoic acid and its positively charged amide analogue, Free Radic Biol Med 26,1418-26. Tirosh, O., Sen, C. K., Roy, S., and Packer, L. (2000). Cellular and mitochondria! changes in glutamate-induced HT4 neuronal cell death, Neuroscience 97, 531-41. Wallach, D. (1997). Apoptosis. Placing death under control, Nature 388,123,125-6. Wei, Y., Cao, X., Ou, Y., Lu, J., Xing, C., and Zheng, R. (2001). SeO(2) induces apoptosis with down-regulation of Bcl-2 and up- regulation of P53 expression in both immortal human hepatic cell line and hepatoma cell line, Mutat Res 490,113-21. Yang, S., Zhu, H., Li, Y., Lin, H., Gabrielson, K., Trush, M. A., and Diehl, A. M. (2000). Mitochondria! adaptations to obesity-related oxidant stress, Arch Biochem Biophys 378, 259-68. Zago, E. B., Castilho, R. F., and Vercesi, A. E. (2000). The redox state of endogenous pyridine nucleotides can determine both the degree of mitochondrial oxidative stress and the solute selectivity of the permeability transition pore, FEBS Lett 478, 29-33. Zhang, Y., and Herman, B. (2002). Ageing and apoptosis, Mech Ageing Dev 123, 245-60. Zoratti, M., and Szabo, I. (1995). The mitochondrial permeability transition, Biochim Biophys Acta 1241, 139-76.

328

Thiol Metabolism and Redox Regulation of Cellular Functions A. Pompella el al. (Eds.) IOS Press, 2002

Redox State of Glutathione and Thioredoxin in Differentiation and Apoptosis Walter H. WATSON, Van CHEN, and Dean P. JONES Department of Biochemistry, Emory University School of Medicine, Atlanta, Georgia, USA 1. Introduction Glutathione (GSH) and thioredoxin (Trx), the two major biological thiol-disulfide redox systems, provide overlapping but complementary functions with respect to detoxification and maintenance of the intracellular redox state. GSH, in conjunction with GSH reductase, GSH-Stransferase, glutaredoxin, and NADPH, provides protection against a wide variety of oxidants and electrophiles. Likewise, Trx provides protection against oxidative stress through its interactions with Trx reductase, peroxiredoxins, and NADPH. GSH is present in millimolar concentrations within cells, and thus has a high capacity for detoxification. Trx, on the other hand, is present in micromolar concentrations, and, therefore, has a lower capacity than GSH for detoxification and repair. However, because Trx contains two cysteines in its active site, Trx is more suited than GSH for 2-electron reduction of proteins, as occurs in the reduction of protein disulfides. Thus, the high capacity of GSH and the efficiency of Trx complement one another in maintenance of thiol-disulfide redox environments. During recent years, considerable evidence has accumulated to show that redox signaling mechanisms function in cell regulation and growth control. Agents altering GSH concentration affect transcription of detoxification enzymes, cell proliferation and apoptosis [1]. In principle, either GSH, GSH S-conjugates, GSSG or the redox state of the GSH/GSSG couple could provide a mechanistic control or signal for functional changes. Both GSH loss and GSH oxidation have been associated with increased expression of the rate-limiting enzyme of GSH synthesis, glutamatexysteine ligase (GLCL), and several other detoxification systems, including glutathione S-transferase (GST) and NAD(P)H: quinone reductase (N:QR). A loss or oxidation of GSH also occurs in association with differentiation both in vitro and in vivo and during apoptosis. In contrast, increases in GSH and/or a reduction of the GSH/GSSG pool are associated with growth stimulation by nutrients and growth factors. Thus, the balance of GSH and GSSG may not only reflect oxidative stress but also may reflect changes in redox signaling and control. Less is known about how the redox state of thioredoxin is maintained and the influence of Trx redox on cellular functions. In vitro assays have clearly shown that the oxidized and reduced forms of Trx have opposite effects on transcription factor activity and regulation of ASK-1. However, there have been very few reports on the redox state of intracellular Trx. In this report, we will discuss our efforts to define the changes in the redox states of glutathione and thioredoxin that occur as cells progress from a proliferating state to a more differentiated state and, finally, to apoptosis. Also, we will discuss the potential relevance of these findings in the regulation of cell signaling and function, as well as in aging and disease.

W.H. Watson ct at. / Redox State of Glutathione and Thioredoxin

329

2. Expression of the Redox State The most widely used indicator of the redox state of the glutathione pool is the ratio of GSH to GSSG. However, for oxidation-reduction reactions involving GSH and GSSG where the reaction is a 2 e" transfer, such as those involving proteins with vicinal thiols, GSSG + Pr-(SH)2 2 GSH + Pr-(SS), the ratio of donor/acceptor is proportional to [GSH]2/[GSSG], and not GSH/GSSG. Because 2 e" transfers appear to be common in redox-sensitive proteins, GSH/GSSG may not be the best measure of redox changes relevant to oxidative stress and redox signaling. A convenient expression for the redox state of the GSH/GSSG pool that incorporates the correct stoichiometry of 2 GSH oxidized per GSSG formed is the redox potential (Eh), calculated according to the Nernst equation: Eh = E0 + (RT/nF)*ln([GSSG]/[GSH]2). In this expression, Eh at a defined pH is the electromotive force given in volts (or millivolts) relative to a standard hydrogen electrode (1 atm Ha, 1 M H+), E0 is the standard potential for the redox couple at the defined pH, R is the gas constant, T is the absolute temperature, F is Faraday's constant and n is the number of electrons transferred. The Eh value provides a quantitative expression of the tendency of the redox couple to accept or donate electrons. It incorporates both an expression of the inherent affinity of the molecule for electrons (E0) and the mass action effect of the concentrations of both the donor and acceptor forms (logarithmic term). Because EH values can be estimated for other redoxactive biomolecules, the parameter provides a convenient way to compare the tendency of the GSH/GSSG pool to donate or accept electrons from other redox-active biologic components. In such processes, couples with a more negative EH value are better reductants. One must bear in mind that EH values calculated as above do not reflect true redox potentials because thioldisulfide components are not in equilibrium in biological systems. Thus, calculated EH values are often described as "redox states" rather than "redox potentials" to emphasize this distinction. 3. Eh of Extracellular GSH/GSSG Studies of GSH/GSSG redox state in human plasma show that this pool is oxidized relative to tissue values [2]. Results indicate that GSH released from tissues reacts with the relatively high concentration of cystine in plasma and thereby helps to maintain the redox state of the cysteine/cystine pool. Oxidative stress in both tissues and in the plasma can be expected to alter the redox state of the plasma pool so that measurement of plasma GSH/GSSG redox state can provide an indicator of in vivo oxidative stress. Accumulating data indicate that extracellular plasma EH is regulated and that perturbation during aging, disease and toxicity may contribute to tissue and organ dysfunction [3,4]. Data are now available on thiol-disulfide redox in plasma [2], the intestinal lumen [5], and cell culture medium [6]. Studies on plasma redox show that the EH values of the low

330

W.H. Watson el at. /Redox State of Glutathione and Thioredoxin

molecular weight thiol/disulfide pools are correlated with each other when compared among different individuals, but the GSH/GSSG couple is 50 mV more reduced than the Cys/CySS pool [2]. Older individuals and diabetics have more oxidized values [3], and values are oxidized following high-dose chemotherapy for bone marrow transplantation [4]. These observations indicate that Eh in plasma varies according to physiology, disease and toxicity. Because there is little variation among young healthy individuals, measures of plasma redox may provide a useful means for clinical assessment of the balance of oxidative stress and opposing defense mechanisms. The consequences, if any, of having a more oxidized plasma redox are not yet clear. Hwang and Sinskey [7] showed in cell culture that cell density varies according to Eh of the culture medium. While their study does not distinguish between effects on cell proliferation and apoptosis, we have found that cell proliferation increases with a more reduced Eh for extracellular Cys/CySS over the range measured in vivo [6]. Apoptosis in tissue culture is inhibited by thiols [8], but whether this effect occurs over the physiologic range of Eh is not known. The quantitatively important mechanisms for maintaining Eh in the plasma have not been experimentally established. Values for the GSH/GSSG pool are considerably more reduced than for the Cys/CySS pool [2], indicating that GSH release and Cys uptake contribute to maintenance of different Eh values for the pools. However, uptake of CySS also occurs in many cell types, and the small intestine regulates extracellular thiol-disulfide redox by a mechanism in which enhanced CySS uptake is associated with stimulated Cys release [5]. This cysteine-cystine shuttle functions in both the lumen and the vascular perfusate [9] to regulate extracellular redox in response to added GSSG. In addition, the basolateral membranes of small intestinal enterocytes and renal proximal tubules also have a thiol oxidase that oxidizes low molecular weight thiols [10]. Whether this system functions along with the transport systems to maintain plasma redox is not known.

4. Eh of IntraceUular GSH/GSSG Eh values for GSH/GSSG in cells and tissues are considerably more reduced than those for extracellular fluids. Essentially all available cellular and tissue values are in the range of-260 to -150 mV, with the more oxidized values only present in cells undergoing apoptosis. The range of values is relatively small considering uiat the NADPH/NADP* value is about -400 mV and that of C«2 as the terminal oxidant is at least +600 mV (Fig.l). The apparent displacement from equilibrium and the relatively small range of values suggest that the redox state of the GSH/GSSG pool is regulated. A provocative aspect of the cellular Eh values is that the cellular Eh tends to vary according to cell growth conditions. Rapidly proliferating cells and tissues have 30 to 60 mV more reduced values than differentiated and growth arrested cells (Fig. 2). Cells undergoing apoptosis are further oxidized by 30 to 60 mV relative to differentiated and growth-arrested cells. These observations indicate that thiol-disulfide redox could provide a context for functional control of cellular processes, essentially providing optimal (or sub-optimal) redox environments for the function of enzymes, transcription factors and other proteins. This could function in much the same way as pH optimum determines the activity of acid phosphatases and alkaline phosphatases in different cellular compartments. An important difference, however, is mat redox changes occur in the cytoplasm so that if specific systems for proliferation, differentiation or apoptosis have different redox optima, then changes in steady-

W.H. Watson et al. / Redox State ofGlutathione and Thioredoxin

Intracellular

-400 -200

Extracellular

NADPH GSH/GSSG

NADH

GSH/GSSG Cys/CySS

> 0 Mitochondria! Electron Transport Chain

^200

400 600 800

Figure 1. Range of biologically relevant redox potentials.

2501

\

O t/5 C/5

I

o

More Reduced

Proliferating Cells

200

Differentiated or Growth Arrested Cells

\ -150J

More Apoptotic Cells Oxidized

Fig. 2. Eh values of the intracellular GSH/GSSG couple become more positive (more oxidizing) as cells progress from a proliferating state to a differentiated or growtharrested state. Cells undergoing apoptosis exhibit the most oxidizing Eh values.

331

332

W.H. Watson et at. / Redox State of Glutathione and Thioredoxin

state cytoplasmic redox could provide a means to optimize conditions for proliferation, differentiation or apoptosis. At present, such a possibility is only hypothetical in that no specific examples of such control have been established. Modeling of redox sensitivities of proteins with E0 values for vicinal dithiols in the range of-200 to -260 mV shows that gains or losses in function of 55fold could occur with a change of 60 mV (Fig. 3). The magnitude of functional change could be amplified by coupling multiple thiol-disulfide couples together or by having redox sensitive groups in chaperones, docking/assembly proteins or nuclear transport systems. Two key conditions would be required for redox-dependent regulation of proteins by GSH/GSSG: enzymes would be required to control protein thiol/disulfide redox in association with changes in GSH/GSSG redox, and metabolic systems would be needed to control GSH/GSSG redox at appropriate values. Enzymes are needed to catalyze exchange between the GSH/GSSG pool and protein dithiol-disulfide motifs because the non-enzymatic exchange rates are too slow to achieve effective regulation under biologic conditions. A protein family, consisting of glutaredoxin and related proteins, catalyzes such exchange reactions [11]. These enzymes are widely distributed and well characterized. However, specific examples of their function in cellular regulation are not well established; this leaves open the possibility that the GSH/GSSG system serves largely in detoxification and has little role in regulation. If so, the GSH/GSSG redox changes associated with proliferation, differentiation and apoptosis may be a secondary indicator of redox control by other thiol-disulfide systems, such as those dependent upon members of the thioredoxin family of proteins.

-260

-240

-220

-200

Eh(mV) Rg. 3. Predicted effect of Eh on enzyme activity. Shown is a hypothetical protein containing a regulatory dithiol with a midpoint potential of -260 mV. Assuming that the reduced (dithiol) form is active, and the oxidized (disulfide) form is inactive, a 60 mV oxidation translates into a 55-fold decrease in the amount of the active form of the protein. The shaded areas highlight the range of GSH/GSSG Eh values measured in proliferating and in differentiated cells, as indicated.

W.H. Watson et al. /Redox State of Glutathione and Thioredoxin

333

5. Redox Reactions involving Thioredoxin Considerable evidence is available to indicate that thioredoxins are required for enzymatic and transcription factor function. In addition, Trx has chaperone-like activity and controls apoptosis signal-regulating kinase-1 (ASK-1) [12] and tumor necrosis factor a-receptor signaling via redox-dependent associations [13]. Thioredoxin functions are distinct from those of GSH/GSSG, and this difference may provide specificity in redox control and signaling. The thioredoxin family of proteins is characterized by the active site sequence WCGPC [14,15]. The midpoint potential (Eo) for this dithiol/disulfide couple is similar in all species examined, ranging from -270 mV in Escherichia coli [16] to -240 mV in yeast [17] and -230 mV in T7 bacteriophage [18]. In humans, there are two distinct forms of Trx: Trxl and Trx2. Trxl is localized predominately in the cytoplasm, but is also found in the nucleus and is exported into the plasma [19]. Trx2 is localized in the mitochondria. Interestingly, mitochondria also contain a distinct form of Trx reductase. The active site cysteines (Cys32 and Cys35 in human Trxl) are readily accessible on the surface of the protein, and become oxidized to a disulfide upon reduction of a target protein (Fig. 4). This disulfide is cycled back to the dithiol by Trx reductase [20]. Unlike Trx2 and Trx's from lower species, mammalian Trxl contains additional conserved cysteine residues (at positions 62, 69, and 73 of human Trx). In X-ray crystal studies, Cys73 was present as an intermolecular disulfide bond (Trx homodimer) [21], suggesting a possible function for Cys73. However, a mutant Trx bearing a serine at this position still appeared as a homodimer in the crystal structure, suggesting that Cys73 was not essential for dimerization [21].

Protein-SS

Trx(SH)2

NADP

Protein(SH)2

Trx-SS

NADPH

Fig. 4. The catalytic cycle of Trx. The active site of Trx becomes oxidized upon reduction of a target protein disulfide (Protein-SS) to a dithiol (Protein(SH)2). The active site is regenerated by the action of Trx reductase, which uses NADPH as the source of electrons.

Oxidized and reduced forms of bovine Trx have been separated by carboxymethylation of thiols, native gel electrophoresis, and immunoblotting [22]. The fully reduced and fully oxidized forms of bovine Trx were identified, but intermediate bands on the immunoblot were only identified as "partially carboxymethylated" by the authors. We have used mass spectrometry to positively identify the forms of human Trx that are resolved by this method (Fig. 5). The results show that there are two redox-active dithiol/disulfide centers in human Trx, one involving Cys32 and Cys35, and the other involving Cys62 and Cys69 [23]. Midpoint potentials (Eo's) for the active site and non-active site oxidations were determined from the

334

W.H. Watson el al. / Redox State of Glutathione and Thioredoxin

relative amounts of each of the redox states of Trx in equilibrium with redox buffers, and found to be -240 mV and -200 mV, respectively. Analysis of redox states in extracts from THP1 cells showed that Trx was maintained at -280 mV and was oxidized in response to toxic concentrations of diamide, a thiol-specific oxidant. Of particular interest, the non-active site disulfide formed only after exposure of the cells to diamide. Cys62 and Cys69 reside in an alpha helix that is proximal to the both the interface domain and the active site of Trx, suggesting that oxidation of the Cys62-Cys69 dithiol may provide a structural switch affecting Trx-protein interactions and Trx function during oxidative stress.

2 Bisulfides 1 Disulfide No Disulfides

Fig. 5. Schematic representation of the Redox Western Wot for Trx1. Lane 1 shows the distribution of fully reduced (no disulfides) Trx1 and partially oxidized (1 disulfide) Trx1 in proliferating cells. Lanes 2 and 3 show the effect of a low (lane 2) and high (lane 3) concentration of diamide on the redox state of Trx1.

In a similar manner, we have used another antibody, specific for Trx2, to assess the redox state of the mitochondria! form of Trx. In proliferating cells, most of the Trx2 was in the reduced form [24]. In cells exposed to diamide, there was a dose-dependent oxidation of Trx2 [24]. Thus, both Trxl and Trx2 were oxidized by diamide. Under appropriate conditions, the Redox Western blot can distinguish between Trxl in the nucleus and Trxl in the cytoplasm [25]. This is a particularly important distinction given the different roles of nuclear and cytoplasmic Trx in the regulation of NFkB activation [26]. The redox state of nuclear Trxl is very similar to that of cytosolic Trxl. In addition, the EH'S of both pools of Trxl are similar to the EH of GSH/GSSG under the same conditions. Exposure of cells to 1 mM tert-butylhydroperoxide (tBOOH) causes a rapid oxidation of nuclear and cytosolic Trxl, as well as cellular GSH/GSSG.

6. Potential for Cellular Regulation via Thioredoxin and Glutathione Redox Proteins with redox-sensitive thiol motifs may be regulated through rapid equilibration with either the GSH/GSSG (via glutaredoxin) or thioredoxin system. Alternatively, regulation could involve transient activation followed by a slower inactivation. Thiols in close proximity to a cationic amino acid are relatively reactive. Under basal conditions, these thiols could

W.H. Watson ct a I. /Redox State of Glutathione and Thioredoxin

335

autooxidize to a stable sulfenic acid or internal disulfide to achieve an OFF state. These proteins could be activated by reduction involving either a thioredoxin or GSH/GSSG system. This could occur following assembly of a complex that allows reduction (ON state). Such a system could have the gain of function (i.e., 55-fold for 60 mV) similar to phosphorylation mechanisms and have the recovery characteristics of many phosphatase-dependent mechanisms (decay to a resting state following activation). Indeed, such a mechanism could control phosphatase activities, as some are thiol-dependent and inhibited by oxidation [27,28]. The second key condition for the function of such a redox control mechanism is the requirement for a means to control GSH/GSSG (and/or thioredoxin) at appropriate redox values. For GSH/GSSG, this control could be regulated simply by the activity of GSSG reductase. Principal control could be exerted by a change in the concentration of the substrate GSSG, and secondary control could involve a change in enzyme expression. However, given the apparent importance of regulating thiol/disulfide redox, it would appear that additional mechanisms would exist. During apoptosis, the redox shift is dependent upon increased generation of reactive oxygen species by mitochondria [29], and mitochondria could also control thiol/disulfide redox during cell proliferation and differentiation by variation in the rate of generation of reactive oxygen species. This control mechanism could involve the known induction of GSH synthesis in response to reactive oxygen species. An increase in pool size would allow cells to maintain a more reduced Eh due to increased GSH and due to an enhanced reduction of GSSG (at a higher steady-state GSSG concentration). Enhanced generation of reactive oxygen species by NADPH oxidases in association with cell proliferation could similarly result in a reduction of thiol/disulfide redox state by increasing the GSH pool size. Alternatively, recruitment of GSSG reductase into a complex could determine its activity in specific pathways. 7. Conclusion The redox of thiol-disulfide components in biological systems is conveniently and simply expressed in terms of an Eh value calculated from the Nernst equation using measured concentrations of reduced and oxidized forms. Eh for cellular glutathione is remarkably constant among different cell types under the same conditions of growth. However, Eh values become progressively more oxidized in cells undergoing proliferation, differentiation and apoptosis. Blood plasma glutathione redox also varies little among young healthy individuals, yet becomes oxidized in association with aging, toxicity and certain diseases, indicating that blood plasma redox measurements may be useful to clinically detect oxidative stress and assess potential interventional strategies. The Eh of thioredoxin (Trxl) can be calculated using the relative amounts of oxidized and reduced forms as measured by the Redox Western blot. Trxl is localized within both the cytoplasm and the nucleus, and, using appropriate precautions, the Redox Western blot can discriminate between redox changes in these two subcellular compartments. More recently, the Redox Western blot approach has been used to assess the redox state of mitochondrial thioredoxin (Trx2). All three pools of cellular thioredoxin (cytoplasmic Trxl, nuclear Trxl, and Trx2) become oxidized in response to peroxides and other oxidants. Thus, like glutathione, Trxl and mtTrx may play fundamental roles in cellular redox regulation.

336

W.H. Watson et al. / Redox State of Glutathione and Thioredoxin

References 1.

W.G. Kiriin, J. Cai, S.A. Thompson, D. Diaz, T.J. Kavanagh, D.P. Jones, Glutathione redox potential in

2.

D.P. Jones, J.L Carlson, V.C. Mody, Jr., J. Cai, M.J. Lynn, P. Stemberg, Jr., Redox state of glutathione in

response to differentiation and enzyme inducers, Free Radio. Biol. Med. 27 (1999) 1208-1218. human plasma, Free Radio. Biol. Med. 28 (2000) 625-635. 3.

P.S. Samiec, C. Drews-Botsch, E.W. Flagg, J.C. Kurtz, P. Stemberg, R.L. Reed, D.P. Jones, Glutathione in human plasma: decline in association with aging, age-related macular degeneration, and diabetes, Free Radical Biol. Med. 24 (1998) 699-704.

4.

C.R. Jonas, A.B. Puckett, D.P. Jones, D.P. Griffith, E.E. Szeszycki, G.F. Bergman, C.E. Furr, C. Tyre, J.L Carlson, J.R. Galloway, J.B. Blumberg and T.R.

5.

Ziegler, Plasma antioxidant status after high-dose chemotherapy: a randomized trial of parenteral nutrition in

5.

bone marrow transplantation patients, Am. J. Clin. Nutr. 72 (2000) 181-189. LJ. Dahm, D.P. Jones, Rat jejunum controls luminal thiol-disulfide redox, J. Nutr. 130 (2000) 2739-2745.

6.

L.T. Miller, J. Cai, W.H. Watson, P. Stemberg, D.P. Jones, Thiol disulfide redox in growth control of HT29 cells, Toxicologist 54 (2000) 211.

7.

C. Hwang, A.J. Sinskey, in: Production of biologicals from animal cells in culture. (R.E. Spier, J.B. Griffiths. B.

8.

D.P. Jones, E. Maellaro, S. Jiang, A.F.G. Slater, S. Orrenius, Effects of N-acetyl-L-cysteine on T-cell apoptosis

Meignier, eds), p. 548-567. Oxford, Halley Court, 1991. are not mediated by increased cellular glutathione, Immunol. Lett. 45 (1995) 205-209. 9.

LJ. Dahm, D.P. Jones, Clearance of glutathione disulfide from rat mesenteric vasculature, ToxkxH. Appl. Pharmacol. 129 (1994) 272-282.

10. L.H. Lash, D.P. Jones, Purification and properties of the membranal thiol oxidase from porcine kidney, Arch. Biochem. Btophys. 247 (1986) 120-130. 11. W.W. Wells, Y. Yang, T.L Deits, Z.R. Gan, Thioltransferases, Adv. Enzymol. Relat. Areas Mol. Biol. 66 (1993) 149-201. 12. M. Saitoh, H. Nishitoh, M. Fujii. K. Takeda, K. Tobiume, Y. Sawada, M. Kawabata. K. Miyazono. H. Ichijo, Mammalian thioredoxin is a direct inhibitor of apoptosis signal-regulating kinase (ASK) 1, EMBO J. 17 (1998) 2596-2606. 13.

H. Liu, H. Nishitoh. H. Ichijo, J.M. Kyriakis, Activation of apoptosis signal-regulating kinase 1 (ASK1) by tumor necrosis factor receptor-associated factor 2 requires prior dissociation of the ASK1 inhibitor thioredoxin, Mol. Cell. Biol. 20 (2000) 2198-208.

14.

H. Eklund, F. Gleason, A. Holmgren, Structural and functional relations among thioredoxins of different

species, Proteins 11 (1991) 13-28. 15. G. Powis, W.R. Montfort, Properties and biological activities of thioredoxins, Annu. Rev. Pharmacol. Toxicol. 41 (2001)261-295. 16. G. Krause, J. Lundstrom, J.L. Barea, C.P. de la Cuesta, A. Holmgren, Mimicking the active site of protein disulfkJe-isomerase by substitution of proline 34 in Escherichia coli thioredoxin, J. Biol. Chem. 266 (1991) 17.

9494-9500. P.G. Porque, A. Baldesten, P. Retchard, Purification of a thioredoxin system from yeast, J. Biol. Chem. 9

18.

(1970)2363-2370. T. Joelson, B-M. Sjoberg, H. Eklund, Modifications of the active center of T4 thioredoxin by site-directed mutagenesis, J. Biol. Chem. 265 (1990) 3183-3188.

19.

T. Tanaka, H. Nakamura, A. Nishiyama, F. Hosoi, H. Masutani, H. Wada, J. Yodoi, Redox regulation by thioredoxin superfamily; protection against oxidative stress and aging. Free Radic. Res. 33 (2000) 851-855.

20. A. Holmgren, Antioxidant function of thioredoxin and glutaredoxin systems. AntbxkJ, Redox Signal. 2 (2000) 811-820. 21. A. Weichsel, J.R. Gasdaska, G. Powis, W.R. Montfort, Crystal structures or reduced, oxidized, and mutated human thioredoxins: evidence for a regulatory homodimer, Structure 4 (1996) 735-751.

W.H. Watson et al. /Redox State of Glutathione and Thioredoxin

337

22. M.R. Fernando, H. Nanri, S. Yoshitake, K. Nagata-Kuno, S. Minakami, Thioredoxin regenerates proteins inactivated by oxidative stress in endothelial cells, Eur. J. Biochem. 209 (1992) 917-22. 23. W.H. Watson, J. Pohl, D.P. Jones, Formation of a non-active site disulfide in human thioredoxin during oxidative stress, Free Rad Biol Med 3 (2001) 863. 24. W.H. Watson, Y. Chen, D.P. Jones, unpublished observations. 25. W.H. Watson, L. Miller and D.P. Jones, Redox Western blot analysis of the redox state of nuclear and cytoplasmic thioredoxin, The Toxicologist 60 (2001) 48. 26. K. Hirota, M. Murata, Y. Sachi, H. Nakamura, J. Takeuchi, K. Mori, J. Yodoi, Distinct roles of thioredoxin in the cytoplasm and in the nucleus, J. Biol. Chem. 274 (1999) 27891-27897. 27. P. Chiarugi, T. Fiaschi, M.L Taddei, D. Talini, E. Giannoni, G. Raugei, G. Ramponi, Two vicinal cysteines confer a peculiar redox regulation to low molecular weight protein tyrosine phosphatase in response to platelet-derived growth factor receptor stimulation, J. Biol. Chem. 276 (2001) 33478-33487. 28. K. Mahadev, A. Zilbering, L. Zhu, B.J. Goldtein, Insulin-stimulated hydrogen peroxide reversibly inhibits protein-tyrosine phosphatase 1b in vivo and enhances the early insulin action cascade, J. Biol. Chem. 276 (2001)21938-21942. 29. J. Cai, D.P. Jones, Superoxide in apoptosis. Mitochondria! generation triggered by cytochrome c loss, J. Biol. Chem. 273 (1998) 11401-11404.

338

Thiol Metabolism and Redox Regulation of Cellular Functions A. Pompella et al. (Eds.) IOS Press, 2002

Redox Regulation of DNA Repair Joseph LUNEC, Marcus S. COOKE and Mark D. EVANS Oxidative Stress Group; Department of Clinical Biochemistry, University of Leicester, Robert Kilpatrick Clinical Sciences Building, Leicester Royal Infirmary NHS Trust, PO Box 65, Leicester, LE2 7LX, UK - Tel: 0044 (0) 116 252 5890, Fax: 116 252 5887 Web: www.le.ac.ub'cb/osK.html - Email: jl20(a)Je.ac.uk

1. Introduction Evolution has selected for sophisticated biochemical mechanisms to maintain a chemically reduced environment in all living cells. Not only is this regulation essential for the numerous biological process within the cell to occur, but it is also a protective mechanism against the many endogenous and exogenous biological oxidations in the cell. However, reactive oxygen species (ROS): which include singlet oxygen ('C^); superoxide ((V ); hydrogen peroxide (H2O2) and lipid hydroperoxides (LOOH) are now increasingly referred to not just as toxic bi-products of oxidative stress, but important signalling molecules (second messengers) for the regulation of gene expression [1]. Excess ROS are harmful because they can cause damage to all classes of macromolecules in particular polyunsaturated lipids, proteins and DNA [2]. In this short chapter we will discuss some of the signal transduction pathways, which can be activated by ROS; the transcriptional regulation of ROS-induced genes, and in particular those which repair DNA damage. Many excellent reviews have been published which explore the mechanisms of ROS signalling pathways, but this article will suggest a novel hypothesis outlining how modulation of redox signalling can control the repair of oxidative damage to DNA.

2. Oxidative DNA Damage DNA is continuously being damaged through interaction with ROS. Such damage may lead to mutation and ultimately to cancer. Oxidative damage to DNA has also been proposed to be involved in diseases such as chronic inflammation, and cardiovascular disease. Oxidative damage to DNA produces multiple base lesions, involving adduction, ring fragmentationand ring saturation. More than twenty separate base lesions have been identified and sophisticated techniques have been developed for determination of such lesions [3]. Relative ease of measurement has pointed toward an increasingly popular marker of in vivo oxidative damage to DNA, 8-oxoguanine (8-oxoG), as the lesion of choice [4]. It can be measured as the base product by either gas chromatography-mass spectrometry (GC-MS) with selected ion monitoring, high performance liquid chromatography (HPLC) with electrochemical detection (EC) and guanase digestion (guanase assay), or as the deoxynucleoside by reversed-phase HPLC-EC (deoxynucleoside assay) [5, 6] or HPLC with tandem mass-spectrometry (HPLC-MS/MS). However, the measurement of this important, mutagenic lesion has been associated with major methodological difficulties. The formation of a European group to improve the quality of its quantitation: the European Standards Committee on Oxidative DNA Damage

./. Lunec et ul. / Redox Regulation of ON A Repair

339

that is being specifically addressed is the baseline level of 8-oxoG in human cells. ESCODD has defined important areas of concern, a) inaccuracies in DMA quantitation; b) poor optimisation of extraction procedures for DNA; c) incomplete, or excess DNA hydrolysis/digestion, and/or artefactual damage induced during sample manipulation.

8-oxoguanine

50.

8-oxoadenine |

I

37.

I

j

25.

VI

§

12.

Baseline

Placebo

Vitamin C

HPLC-EC (8-oxoG base)

b) O

B, c«

§

Baseline

Placebo

Vitamin C

Timescale (weeks) Fig. 1. The effects of supplementation of 500mg/day vitamin C on thirty normal subjects upon oxidative DNA damage. The 8-oxo(d)G was measured by three separate techniques (a) GC-MS, (b) HPLC-EC (guanase), (c) HPLC-EC (deoxynucleoside), because of the controversy surrounding the quantification of this lesion.

340

J. Lunec et al. /Redox Regulation ofDNA Repair

ESCODD established in its first report the importance of appropriate standardisation and quality control material particularly for GC-MS [6]. In a later report it was confirmed, unequivocally, that GC-MS was overestimating the level of S^oxoG by two to three times the target value within a synthetic oligomer [7]. The utility of 8-oxoG measurements in peripheral blood mononuclear cells (PBMC) is illustrated by an ascorbic acid intervention study, in which 8-oxo(d)G was measured by three different methods (Figures la-c). The level of 8-oxoG was of the order of 2-3 times higher when measured by GC-MS and guanase assays, compared to the deoxynucleoside assay. The GC-MS and guanase assays share a methodological similarity, formic acid hydrolysis, which differs from the deoxynucleoside assay in which enzymatic hydrolysis is used. This difference in methodology may account for the higher levels of 8-oxoG seen with the former procedures [7j. Furthermore, the GC-MS method showed, on average a 20% higher level of 8-oxoG compared to the guanase assay (Figures la&b), which might be accounted for by artefactual production of 8-oxoG during derivatisation. Despite differences in absolute quantitation the figure illustrates quite clearly that using three separate methods for analysing the DNA samples the effect of vitamin C supplementation is to always reduce 8oxo(d)G. 3. Urine and Serum Levels of 8-oxodG In the ascorbic acid intervention study described [5], levels of PBMC 8-oxoG showed a significant negative correlation (p<0.0001) with plasma ascorbic acid levels, which was mirrored by a significant positive correlation between 8-oxoadenine (8-oxoA) and ascorbic acid (p<0.0001). An albeit less significant correlation (p<0.04), was also seen between serum 8-oxodG, as determined by ELISA, and plasma ascorbic acid. A highly significant (p<0.001) increase in serum 8-oxodG was observed after 6 weeks of ascorbic acid supplementation, which returned approximately to baseline levels following washout (Figure 2a). Although longitudinal profiles for urinary excretion of 8-oxodG were similar they showed no such correlation with ascorbic acid (Figure 2b). This was probably due to a lag-time between elimination of 8-oxodG from the cell and its appearance in serum and then subsequently urine. Comparison between basal levels of urinary 8-oxodG by ELISA, and those quantitated by HPLC-EC reveal the former to be higher, although in strong agreement with other groups also using antibody technology. An explanation for the higher levels could be that the antibody-based method does not distinguish between free 8-oxodG and 8-oxodG-containing oligonucleotides. Indeed, we have recently shown in our laboratory that the antibody used in this study also recognises 8-oxodG located within a random 20-mer oligonucleotide in addition to free 8-oxodG (Cooke et al. unpublished observations). The oxidatively modified base is not detected by the antibody. We have shown, by three different methods, that 500mg/day vitamin C significantly reduces PBMC levels of 8-oxodG, irrespective of baseline values. Furthermore, similar results have now been shown by several different groups. Halliwell et al., have in fact reproduced similar changes in the levels of 8-oxoG and 8-oxoA following 6 weeks of vitamin C and iron cosupplementation [8]. A similar, initial increase in 8-oxoA was seen in PBMC following administration of the plant antioxidant lycopene to human volunteers [9]. A novel mechanism of action is therefore required to explain the increase in 8-oxoA corresponding with a decrease in 8-oxodG in PBMC reflected by an increase in serum and urinary 8oxodG.

J. Luncc et al. / Redox Regulation of DNA Repair

341

4. Enzymic Repair of 8-oxodG DNA repair plays a very important part in. maintaining genomic integrity and deficiencies in repair enzyme systems are known to promote cancer development. Since 8-oxoG is probably the most prevalent lesion generated by the action of ROS on DNA, it is perhaps not surprising that a specific enzyme exists for its rapid repair [10, 11]. It appears that 8oxoG is predominantly repaired by a glycosylase, termed hOggl (human 8-oxoguanine glycosylase). The repair mechanism involves the hydrolysis of the N-glycosyl bond of 8oxodG, resulting in the removal of the damaged base [12]. This is rapidly followed by an additional lyase activity towards the apurinic site leaving a strand break. A hOgg2 enzyme has also been defined in human cells as removing 8-oxoG when derived from the nucleotide pool and misincorporated opposite G or A. Other, repair pathways include: (i) nucleotide excision repair (NER), which eliminates 8-oxodG by excising an oligomer by the co-ordinated action of several separate enzymes; (ii) mismatch repair, involving the enzyme human MutY homologue (MYH) which removes adenine or guanine when mismatched opposite 8-oxoG, repair resynthesis will be more likely to incorporate cytosine opposite 8-oxoG, enable hOggl another chance to repair 8-oxoG; (iii) prevention of incorporation is executed by 8-oxo-2' deoxyguanosine 5'-triphosphate pyrophosphohydrolase (8-oxodGTPase, hMTHl). Although apparently minor pathways of repair of 8-oxodG only NER (or transcription-coupled repair) or hMTHl activity could result in 8-oxodG excretion [10, 11].

5. Redox Regulation of Gene Expression Dietary antioxidants exert their protective effects either through scavenging of ROS or by stimulating endogenous defenses via signal transduction pathways. Some enzymes involved in cellular defense are controlled at the transcriptional level by the antioxidant responsive element (ARE) or activator protein-1 (AP-1) sites found in the promoter region of their genes. At least two major transcription factors, AP-1 and NFicB appear to be regulated by intracellular redox status [13]. The mechanisms related to how both NFicB and AP-1 bind to DNA and result in transcription of genes is well covered by many reviews and will not be detailed here. However, it is important to state that although oxidative stress is one mechanism which can activate these transcription factors, a wide range of other stimuli have been identified. These stimuli include UV and y radiation, tumour necrosis factor-oc (TNF-a); interleukin-1 (IL-1), phorbol myristate acetate (PMA) and paradoxically some antioxidants. Although these represent a very diverse set of stimuli, all involve second messenger pathways that may cause protein phosphorylation. In particular, such activation can lead in particular to regulation of cytokines; chemotactic factors, adhesion molecules and possibly DNA repair genes. Both hOggl and human apurinic/apyrimidinic endonuclease (hAPE) [14] have an ARE binding site in their promoter region as well as AP-1 binding sites suggesting both genes are redox sensitive. It is well established that hAPE is inducible by oxidative stress and has multifunctional properties: acting as a DNA repair enzyme, eliminating abasic sites in DNA and orchestrating gap filling by DNA polymerase; it appears to enhance the repair activity of hOggl; and also has the capacity to restore DNA binding to Fos-Jun heterodimers and JunJun homodimers. This latter action is via the cysteine residue on the N-terminus of the hAPE, occuring in conditions of oxidative stress (reducing factor/Ref-1 activity). It is possible that hAPE may also regulate p53 by converting the inert form to the active form, therefore arresting cell cycle events [15].

J. Lunec et al. / Redox Regulation of DNA Repair

342

o T3 O X O

00

s O»

WeekO

Baseline

Week 3

Week 6

Placebo

Week 9

Week 12

Vitamin C

Weak

Week 25

Washout

o

I "o ^o.

.c 'c e5

£w

o •o o o

Week 3

Baseline

Week 6

Placebo

Week 9

Week 12

Vitamin C

Week 19

Washout

Time course (weeks) Fig. 2. The effect of supplementation of 500 mg/day vitamin C on thirty normal subjects with respect to repair of the oxidative DNA lesion, 8-oxodG, measured in (a) serum and (b) urine. Asterisks indicate a significant (*** p<0.001) difference from baseline and placebo measurements.

J. Lunec et al. / Redox Regulation of DNA Repair

343

6. Ascorbic Acid and DNA Repair Our initial analysis of 8-oxoG levels in PBMC (figure 1) suggested that ascorbic acid might indeed be acting directly as an antioxidant in vivo because levels were reduced upon supplementation and returned to baseline on removal of the ascorbic acid. However, the simultaneous increase in 8-oxoA observed also suggested a concomitant pro-oxidant effect. The observation that 8-oxodG was released into the serum (possibly in the form of oligonucleotide fragments) and subsequently increased in urine, indicated that the action of ascorbic acid in reducing 8-oxodG levels was not due to inhibition of 8-oxoG formation, but rather promotion of its removal (Figure 2a & 2b) [10]. There are at least two possible sources of 8-oxodG in urine; first, is through cell turnover and DNA degradation; second, through repair of DNA/the nucleotide pool. To account for the increase in serum/urinary 8oxodG, ascorbic acid could influence a number of processes. Ascorbic acid may: (i) act as a pro-oxidant for guanine moieties not contained in DNA e.g. dGTP giving rise to 8oxodGTP, ultimately yielding 8-oxodG; (ii) promote the repair and/or purging of 8-oxodG from the nucleotide pool/DNA [11]. The timing of the appearance of 8-oxodG at maximum levels in urine (Figure 2) compared to serum and PBMC would suggest ascorbic acid is having a residual effect, detectable long after plasma values have returned to baseline. It may be that the processing of such lesions accounts for the delay between removal of the lesion from DNA and its appearance in the urine. Account can now, be taken of the observed simultaneous increase in 8-oxoA. The enzyme hOggl removes only the base, however, our results in serum and urine suggest removal of the deoxynucleoside since the antibody used measures only the deoxynucleoside and not the base. The increase in 8-oxoA could only be explained if ascorbic acid initially increases both lesions, via a pro-oxidant activity priming an adaptive response which up-regulates DNA repair processes, which in turn specifically eliminates 8-oxodG. The pro-oxidant nature of ascorbic is illustrated in CCRF cells (Figure 3) following one hour of exposure to ascorbic acid at a final concentration of 100 uM. This concentration mimics that surrounding PBMC following supplementation of volunteers with 500 mg ascorbic acid for six weeks.

Fig. 3. The effect of adding vitamin C at a final concentration of 100 jjM on peroxide formation in CCRF cells. Peroxides were detected with the reagent dichlorodihydrofluoroscein, which emits green fluorescence in the presence of peroxides. A is control and B is treated.

J. Lunec el al. / Redox Regulation of DNA Repair

344

[Vitamin C]

2

1 < •«-* «_!

(A)

IR 2I £I

S 1 w S

- s

AP-1

I I (B)

3 «8

2w a

-

a

§

[Vitamin C] Fig. 4. In vitro activation of AP-1 DNA binding by addition and incubation of ascorbic acid with CCRF cells, showing a dose response up to levels present in serum or plasma following supplementation. Doses higher than 100 nM inhibit binding. No cell death was recorded up to 500 \M ascorbate.

7. Redox Regulation of AP-1

AP-1 is an inducible transcription factor containing the protein products fos and jun. The DNA binding activity of fos and jun is regulated in vitro by a redox-dependent, posttranslational mechanism. A conserved cysteine is the focus for both oxidation and

./. Lunec et at / Redox Regulation of DNA Repair

345

reduction leading to DNA binding. Under antioxidant conditions strong DNA binding and transactivation of genes regulated by AP-1 is observed, whereas under pro-oxidant conditions there is only weak DNA binding of AP-1 and weak transactivation of genes [13]. In Figure 4 we show the effect ascorbic acid on AP-1 DNA binding following incubation of CCRF cells with physiologically relevant concentrations of ascorbic acid. Binding was optimal after 3 hours incubation (three fold increase), with 50-100 jiM ascorbic acid. This coincides with the known extra-cellular concentration of ascorbic acid in individuals undergoing supplementation (500 mg/day) with the vitamin over a period of 6 weeks. Is this increased binding of AP-1 to DNA due to the pro-oxidant or anti-oxidant effect of ascorbic acid? It appears that formation of a heterodimer of c-fos and c-jun proteins or a cJun homodimer is involved in promoting API-DNA binding. Transcription of the c-fos promoter is regulated by serum response factor SRF/TCF and the cyclic AMP-responsive factor CREB, while AP-1 activates c-jun promoter as a positive auto-regulator [14]. Our observations on AP-1 activation and binding suggest that c-jun mRNA is induced relative to c-fos as indicated by the incubation of cells with antibodies to both jun and fos (unpublished observations) leading to weak DNA binding and weak transactivation. Under conditions of incubation with 50 uM ascorbic acid the c-jun mRNA is induced prior to cfos while upon antioxidant stimulation (100 uM) c-fos mRNA is induced first followed by induction of c-jun. In vitamin C stimulated CCRF cells newly synthesised c-fos may interact with pre-existing c-jun molecules to form AP-1 which in turn activates the c-jun promoter by binding to its AP-1 site [15]. AP-1 controls the expression of many genes, including those encoding collagenase, stromelysin, cyclin D, TGFlp and many cytokines.

8. The Important Role of Glutathione ROS dependent redox cycling of cysteinyl thiols is critical for the interactions between proteins and DNA, which determine signal transduction pathways. It immediately follows therefore that glutathione (GSH) is a key regulator of the redox state of the protein cysteinyl thiols. Due to the low concentrations of GSSG (oxidised glutathione) relative to GSH in the cell, small increases in ROS will produce large increases in GSSG. This in turn will promote oxidation of protein cysteinyl thiols. The resulting shift in equilibrium of thiol disulphide-exchange will effect protein conformation. Reduction of linked disulphides and reversion to previous conformation, is often mediated by reductants such as thioredoxin, glutaredoxin and protein disulphide isomerases [16]. Cycling of cysteinyl residues by redox mechanisms regulates the activity of many transcription factors including AP-1, MAF and NRL [17]. Other important signal transduction molecules such as protein kinase C, collagenase and tyrosine kinases are also affected by redox status. Expression appears to regulate binding of AP-1 to transcription response elements (TRE) in the promoters of these genes. Under oxidant conditions critical cysteine residues of Fos/Jun are modified to sulphenic and sulphonic acid derivatives. Ref-1, which regulates binding by reducing AP1, is identical to hAPE which makes it a link between oxidative stress, redox regulation of transcription and DNA repair.

9. AP-1 Binding and DNA Repair Ref-1 (alternatively referred to as hAPE) is a bifunctional protein, which in addition to functioning as a DNA repair protein, is an important component of the signal transduction processes that regulate eukaryotic gene expression in response to cellular stress (it is itself

J. Lunec et al. /Redox Regulation ofDNA Repair

346

induced by oxidative stress). Thioredoxin maintains Ref-1 in a reduced state thereby facilitating its interaction with fos and jun elements of AP-1 binding and promoting AP-1 binding to DNA [15-17]. Our most recent results show that vitamin C induces de novo synthesis of fos and predominantly jun, presumably through the redox activation of AP-1 binding. Important sequelae of this may be the induction of p53, cell cycle arrest and repair of 8 oxodG, but not 8 oxodA, through a pro-oxidant action. The induction of nucleotide excision repair by this route would be entirely consistent with previous literature indicating links between AP-1 p53, ERCC-1 and nucleotide excision repair [18, 19].

AA

Fig. 5. Scheme illustrating the mechanism by which vitamin C, through redox pathways, may ultimately affect DNA repair. (AA, ascorbic acid; DMA, dehydroascorbic acid; LOO*, lipid peroxyl radical; NER, nucleotide excision repair, TCR, transcription coupled repair.)

10. Conclusions The results reported herein describe a complex effect with 8-oxodG being decreased by intervention, while 8-oxoA was increased significantly, over the six weeks of supplementation. This is entirely consistent with similar studies measuring oxidative DNA damage in vivo following supplementation with antioxidants [18, 19]. Our conclusion was that ascorbic acid had what appeared to be a profound antioxidant effect in vivo (exemplified by a reduction in 8-oxodG), paradoxically balanced by a pro-oxidant effect on adenine. This result was difficult to explain in terms of a putative antioxidant effect of ascorbic acid. We have subsequently discovered that elimination and excretion of 8-oxodG can be followed through DNA, serum and urine measurements and that this may be an

J. Lunec et al. / Redox Regulation of DNA Repair

347

indication of the in vivo repair of this lesion [20]. This strongly suggests that the apparent antioxidant effect of ascorbic acid may initially have been mistaken for a classical scavenging activity, whereas up-regulation, or priming of the-enzyme(s) which eliminate-8oxodG, may in fact have occurred. If this hypothesis is correct it could explain how levels of 8-oxoG decreased, while 8-oxoA values increased in our supplementation study. Ascorbic acid can act as a pro-oxidant in vitro and may initiate an adaptive response, which has been implicated in previous intervention studies associated with the protective effect of antioxidant rich diets [20, 21]. Such a response is likely to be influenced by transcription factors, some of which are in turn dependent on the redox status of the cell [22]. These novel findings challenge our fundamental ideas about natural "so-called" antioxidants in the human diet and define alternative explanations for their protective effects against DNA damage and ultimately carcinogenesis..

Acknowledgements J. Lunec would like to thank The Food Standards Agency and The Scottish Office for financial support of the Oxidative Stress Group.

References [I] Daiton T.P., Shertzer H.G. and Puga A. (1999). Regulation of gene expression by reactive oxygen. Ann. Rev. Pharmacol. Toxicol. 39: 67-101. [2] Halliwell, B. Current Status Review, free radicals, reactive oxygen species and human disease, a critical evaluation with special reference to atherosclerosis. British J. Exp. Path. 70: 737-757; 1989. [3] Dizdaroglu, M. Chemical determination of free radical-induced damage to DNA. Free Rad. Biol. Med. 10: 225-242; 1991. [4] Kasai, K. Analysis of a form of oxidative DNA damage, 8-hydroxy-2'-deoxyguanosine, as a marker of cellular oxidative stress during carcinogenesis. Mutat Res. 387:147-163; 1997. [5] Podmore I.D., Griffiths H.R., Herbert K.E., Mistry N., Mistry P., Lunec J. (1998) Vitamin C exhibits prooxidant properties. Nature 392: 559. [6] Lunec J. (1998) ESCODD: European Standards Committee on Oxidative DNA Damage. Free Rad . Res. 29: 601-608. [7] Lunec J., Herbert K.E., Jones G.D.D., Dickenson L, Evans M.D., Mistry N.. Chauhan D., Capper G., Zheng Q. (2000). Development of a quality control material for the measurement of 8-oxo-7,8 dihydro-2deoxyguanosine, an in vivo marker of oxidative stress, and comparison of results from different laboratories. Free Rad. Res. 33: 521-526. [8] Rehman A., Colli S.C., Yang M., Kelly M., Diplock A.T., Halliwell B., and Rice-Evans C. (1998) The effects of co-supplementation of iron and vitamin C on oxidative damage to DNA in healthy volunteers, Biochem, Biophys. Res. Commun. 246: 293-298. [9] Rehman A., Bourne L.C., Halliwell B., Rice Evans C. (1999) Tomato consumption modulates oxidative DNA damage in humans. Biochems. Biophys. Res. Comun. 262: 828-831. [10] Cooke M.S., Evans M.D., Podmore I.D., Herbert K.E., Mistry P., Hickenbotham P.T., Husseini A., Griffiths H,R,, Lunec J, (1998) Novel repair action of vitamin C upon in vivo oxidative DNA damage, Feb$ Letts. 439: 363-367. [II] Cooke M.S., Evans M.D., Herbert K.E., Lunec J. (2000) Urinary 8-oxo2' deoxyguanosine. source, significance and supplements. Free Rad. Res. 32: 381-397. [12] Arai K., Morishita K., Shinmura K., Kohno T., Kim S.R., Nohmi, Taniwaki M., Ohwada S., Yokata J. (1997), Cloning of human homolog of the yeast OGG1 gene that is involved in the repair of oxidative DNA damage. Oncogene 14: 2857-2861. [13] Meyer M., Pahl H.L., Baeuerie P.A. (1994) Regulation of the transcription factors NFKB and AP-1 by redox changes. Chemico-Biologicai interactions, 91: 91-100. [14] Groseh S., Kaina B. (1999) Transcription activation of Apurinic/Apyrinidimic Endonuclease (Ape, Ref-1) by oxidative stress requires CREB, Biochem. Biophys. Res. Commun. 261: 859-863. [15] Xanthoudakis S., Curran T., (1996) Redox regulation of AP-1. A link between transcription factor signalling and DNA repair. In: Biological Reactive Intermediates V Ed by Snyder R. Plenum Press New York, pp69-75. [16] Pan/is G., Mustachich D., Coon A. (2000) The role of the redox protein thioredoxin in cell growth and cancer. Free Rad. Biol. Med. 29: 312-322.

348

J. Lunec et al. / Redox Regulation ofDNA Repair

[17] Schenk H., Klein M., Edbrugger W., Droge W., Schufee-Ostholt K. (1994) Distinct effects of thioredoxin and antioxidants on the activation of transcription factors NF-KB and AP-1. Proc. Nat). Acad. Sci. USA 91: 1672-1676. [18] Li Q., Tsang B., Bostick-Bruton F., Reed E. (1999) Modulation of excision repair cross complementation group 1 (ERCC-1) mRNA expression by pharmacological agents in human ovarian carcinoma cells. Biochem Pharmacol. 57: 347-353. [19] Kirch H.C., Flasivinkels S., Rump F.H. , Brockmann D.. Esche H. (1999) Expression of human p53 requires synergistic activation of transcription from the p53 promoter by AP-1. NF-kappa B and myc/max Oncogene 18: 2728-2738. [20] Kaina B.. Kiesel S., Grombacher T.. Fritz G. (1998) Inducible protective functions of mammalian cells against DNA damaging agents. Periodicum Biologorum 100: 339-344. [21] Ramana C.V., Boldogh I., Iziumi T., Mitra S. (1998) Activation of apurinic apyrimidimic endonudease in human cells by reactive oxygen species and its correlation with their adaptive response to genotoxicity Proc. Natl. Acad. Sci. USA. 95: [22] Adler V., Zhimin Y., Tew K.O.. Ronai Z. (1999) Role of redox potential and reactive oxygen species in stress signalling. Oncogene 18: 6104-6111.

349

Author Index Aronis, A. Asare, N. Balcerczyk, A. Bánhegyi, G. Barber, V.S. Bartosz, G. Bast, A. Benedetti, A. Biesalski, H.K. Bilinski, T. Bluvshtein, E. Brigelius-Flohé, R. Budde, H. Cappiello, M. Cecconi, I. Chen, Y. Coles, B.F. Comporti, M. Cooke, M.S. Cotgreave, LA. Csala, M. Csermely, P. Dal Monte, M. Del Corso, A. Dicato, M. Dickinson, D.A. Diederich, M. Dominici, S. Duvoix, A. Emdin, M. Enoiu, M. Evans, M.D. Flohé, L. Forman, H.J. Frank, J. Fulceri, R. Galteau, M.-M. Griffiths, H.R. Grzelak, A. Habib, G.M. Haenen. G.R.M.M. Herber, R. Hogg, P.J. Homolya, L. Hughey, R.P.

317 172 75 22,29,38 238 75 230 29,38 252 75 160 96 85 299 299 328 119 209 338 290 29,38 273,281 299 299 138 1 138 209 138 223 197 338 85,96 1,306 252 29 138 238 75 14 230 197 265 107 146

Huseby, N.-E. Jakubowski, W. Janaszewska, A. Jean, J.-C. Jiang, X.-M. Jones, D.P. Joyce-Brady, M. Kadlubar, F.F. Komlosh, A. Korcsmáros, T. Koziol, S. Leroy, P. Lieberman, M.W. Liu, Y. Lorenzini, E. Lunec, J. Maellaro, E. Mandl, J. Marc, R.E. Marszalek, M. Matthias, L.J. Mikkelsen, I.M. Minotti, G. Morceau, F. Mortensen, B. Mura, U. Nardai, G. Nöhammer, G. Paolicchi, A. Passino, C. Paules, R.S. Pieri, L. Pompella, A. Porat, N. Rojas, E. Rychlik, B. Sarkadi, B. Schmitz, M. Schnekenburger, M. Sen, C.K. Shi, Z.-Z. Shilo, S. Slyshenkov, V.S. Soszynski, M. Soti, Cs.

172 75 75 182 265 328 182 119 160 281 75 197 14 182 209 338 209 22,29,38 182 75 265 172 209 138 172 299 273,281 48 209,223 223 14 209 209,223 160 14 75 107 138 138 317 14 317 61 75 273

350

Stark, A.-A. Steinberg, P. Tirosh, O. Torres, M. Tubi, C. Valverde, M. Váradi, A.

160 160 317 306 160 14 107

Vilardo, P.O. Volohonsky, G. Watson, W.H. Wellman, M. Wetting, S. Wojtczak, L. Yam, P.T.W.

299 160 328 172,197 172 61 265